**Prosthetic Shoulder Joint Infection by** *Cutibacterium acnes***: Does Rifampin Improve Prognosis? A Retrospective, Multicenter, Observational Study**

**Helem H. Vilchez 1,\* , Rosa Escudero-Sanchez <sup>2</sup> , Marta Fernandez-Sampedro 3, Oscar Murillo 4, Álvaro Auñón 5, Dolors Rodríguez-Pardo <sup>6</sup> , Alfredo Jover-Sáenz <sup>7</sup> , M<sup>a</sup> Dolores del Toro <sup>8</sup> , Alicia Rico <sup>9</sup> , Luis Falgueras 10, Julia Praena-Segovia 11, Laura Guío 12, José A. Iribarren 13, Jaime Lora-Tamayo 14, Natividad Benito <sup>15</sup> , Laura Morata 16, Antonio Ramirez 17, Melchor Riera 1, Study Group on Osteoarticular Infections (GEIO) † and the Spanish Network for Research in Infectious Pathology (REIPI) †**


**Abstract:** This retrospective, multicenter observational study aimed to describe the outcomes of surgical and medical treatment of *C. acnes*-related prosthetic joint infection (PJI) and the potential benefit of rifampin-based therapies. Patients with *C. acnes*-related PJI who were diagnosed and treated between January 2003 and December 2016 were included. We analyzed 44 patients with *C. acnes*-related PJI (median age, 67.5 years (IQR, 57.3–75.8)); 75% were men. The majority (61.4%) had late chronic infection according to the Tsukayama classification. All patients received surgical treatment, and most antibiotic regimens (43.2%) included β-lactam. Thirty-four patients (87.17%) were cured; five showed relapse. The final outcome (cure vs. relapse) showed a nonsignificant

**Citation:** Vilchez, H.H.; Escudero-Sanchez, R.; Fernandez-Sampedro, M.; Murillo, O.; Auñón, Á.; Rodríguez-Pardo, D.; Jover-Sáenz, A.; del Toro, M.D.; Rico, A.; Falgueras, L.; et al. Prosthetic Shoulder Joint Infection by *Cutibacterium acnes*: Does Rifampin Improve Prognosis? A Retrospective, Multicenter, Observational Study. *Antibiotics* **2021**, *10*, 475. https://doi.org/10.3390/ antibiotics10050475

Academic Editor: Marc Maresca

Received: 28 March 2021 Accepted: 19 April 2021 Published: 21 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

trend toward higher failure frequency among patients with previous prosthesis (OR: 6.89; 95% CI: 0.80–58.90) or prior surgery and infection (OR: 10.67; 95% IC: 1.08–105.28) in the same joint. Patients treated with clindamycin alone had a higher recurrence rate (40.0% vs. 8.8%). Rifampin treatment did not decrease recurrence in patients treated with β-lactams. Prior prosthesis, surgery, or infection in the same joint might be related to recurrence, and rifampin-based combinations do not seem to improve prognosis. Debridement and implant retention appear a safe option for surgical treatment of early PJI.

**Keywords:** *Cutibacterium acnes*; prosthetic joint infection; surgical and medical treatment

#### **1. Introduction**

*Cutibacterium* (formerly known as *Propionibacterium*) *acnes* is an anaerobic Grampositive bacillus and a skin commensal organism with a predilection for pilosebaceous follicles, and it was formerly considered a contaminant. Moreover, *C. acnes* has been identified as a cause of biomaterial-related infections (BRIs) involving arthroplasty, cerebrospinal fluid (CSF) shunts, and spinal instrumentation, among others [1–3]. In recent years, with improved diagnosis methodology, including prolonged incubation protocols, *C. acnes* has become the microorganism most frequently related to infections involving shoulder prostheses. This infection type has become an emerging problem, but the relevant data are still limited [1,4,5].

*Cutibacterium* infections are usually characterized by a paucity of classical infections or inflammation symptoms, and they are often characterized by the absence of elevated inflammatory markers [1,6].

The role of *C. acnes* in prosthetic joint infections (PJIs) might be underestimated for the following reasons: (1) it is a common contaminant of the skin; (2) it needs a special transport medium; (3) it has delayed growth (up to 14 days); (4) the cultures need to be rechecked or discarded within 3 to 5 days of incubation. The advent of matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) for the routine diagnosis of bacterial infections in clinical laboratories has increased the speed and ease of anaerobic bacteria identification [4,7,8].

*Cutibacterium* appears to have a greater predilection for infections involving the shoulder joint compared to other anatomical regions. The risk factors for *C. acnes-*related orthopedic infection include a history of joint surgery prior to the index surgery and male sex [9,10].

*C. acnes* is usually susceptible to a wide range of common antibiotics but there are no clinical trials or extensive observational studies that allow us to know the best antibiotic regimen or surgical procedure in these patients. The Infectious Diseases Society of America (IDSA) guidelines recommend penicillin or ceftriaxone as first-line treatment for *C. acnes*-related PJIs, with clindamycin or vancomycin as alternatives, and minocycline or doxycycline for suppressive therapy [11]. However, there have also been reports of increased antimicrobial resistance in biofilm-associated *C. acnes* isolates in vitro. In vitro and animal models of *C. acnes* biofilms suggest the efficacy of rifampin against *C. acnes*related foreign-body infections [12,13], but adjunctive rifampin therapy is not included in the IDSA recommendations for *C. acnes-*related PJI management.

Despite its antimicrobial susceptibility, *C. acnes* is sometimes remarkably difficult to eradicate; therefore, medical management of PJIs without surgical intervention has been considered to result in poorer clinical outcomes [2].

The aim of this study was to describe the epidemiological, clinical, and biological characteristics, as well as the outcomes of surgical and medical treatment, of *C. acnes*-related PJI and the potential benefit of rifampin-based therapeutic combinations.

#### **2. Results**

Forty-six cases of *C. acnes*-related PJI were identified, of which two patients were excluded because both had co-infections with a microorganism other than CNS. Finally, we included 44 patients with *C. acnes*-related PJI. The median patient age was 67.5 years (IQR, 57.3–75.8); 75% of the patients were men. The number of cases included, according to year, is shown in Figure 1.

**Figure 1.** Cases frequency by year.

#### *2.1. Patient Baseline and Clinical Characteristics*

Demographic data, comorbidities, risk factors predisposing to PJI, signs and symptoms, and laboratory data at presentation are shown in Table 1. Most cases were classified as late chronic infection (type 2) or positive intraoperative culture (type 4), with 25% being acute prosthetic infections according to the Tsukayama classification. However, according to the Zimmerli classification, the most frequent type of infection was early infection (52.3%), while delayed and late infections were present in 47.7% of cases.

**Table 1.** Demographic, clinical, and laboratory characteristics of shoulder PJI due to *C. acnes.*



Abbreviations: CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell. <sup>a</sup> Data are the number (%) of cases. <sup>b</sup> Median (IQR, interquartile ranges).

#### *2.2. Microbiological Characteristics and Antimicrobial Susceptibility Patterns*

With regard to microbiological data, diagnosis was performed preoperatively and/or intraoperatively in all patients. In 17 (38.6%) of the 44 patients, *C. acnes* was found in the joint fluid aspiration. In 42 (95.5%) of 44 patients, *C. acnes* was found in intraoperative samples. There were 15 patients with *C. acnes* isolation in both samples (joint and intraoperative).

Three or more positive cultures were obtained in 32 patients (72.7%), two cultures were obtained in seven patients (15.9%), and only one culture was obtained in five patients (11.4%), where the infection was demonstrated by histopathologic inflammation and positive sonicate fluid from the prosthetic material culture. All tested isolates were susceptible to β-lactams (penicillin), vancomycin, and rifampin (Table 2).

**Table 2.** Samples and microbiological characteristics of shoulder PJI due to *C. acnes.*


Abbreviations: CLSI, Clinical and Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial susceptibility testing. <sup>a</sup> Data are the number (%) of cases; susceptibilities determined as per CLSI/EUCAST breakpoints. <sup>b</sup> All antibiotics were not tested in all strains isolated.

#### *2.3. Surgical and Medical Therapy*

All patients received surgical treatment: two-stage procedure (38.6%), debridement and implant retention (DAIR) (36.4%), one-stage procedure (18.2%), arthrodesis (2.3%), and resection arthroplasty (4.5%). When we compared the surgical treatment received with the type of infection according to the Tsukayama classification, there was an expected association between performing DAIR and early postoperative infection (Table 3).


**Table 3.** Comparison between the types of treatment with type of infection of shoulder PJI due to *C. acnes.*

<sup>a</sup> Data are the number (%) of cases. <sup>b</sup> Tsukayama classification: early postoperative infection (Type 1), late chronic infection (Type 2), and positive intraoperative cultures (Type 4).

> The majority (43.2%) of antibiotic regimens used β-lactam (amoxicillin), while clindamycin was used in 31.8% and other antibiotics (linezolid, quinolones, doxycycline, and glycopeptides—vancomycin and teicoplanin) were used in 22.7%. Rifampin was administered concurrently with at least one of the aforementioned antibiotics in 19 patients (43%), with two cases of rifampin treatment being discontinued due to adverse reactions. When we compared the type of antibiotic treatment with the type of infection, we observed no significant differences (Table 3). The median duration of antibiotic therapy was 56 days (IQR, 44–84 days).

#### *2.4. Treatment Outcomes*

Among the 44 patients included, 39 were evaluable for treatment outcome. At the last follow-up, five patients were lost, 34 patients were considered cured, and five had microbiologically confirmed recurrence. Three patients died due to noninfectious causes (acute pulmonary edema, advanced renal neoplasm, and cardiorespiratory arrest); these patients were followed up for more than 12 months with favorable infection outcomes.

We compared patients with a favorable outcome to those who failed treatment (Table 4). All patients in the failure group were male, but there was no significant difference in the clinical presentation, treatment received, or type of infection. A nonsignificant trend toward a higher frequency of failure was observed among patients with previous prosthesis (odds ratio (OR): 6.89; 95% confidence interval (CI): 0.80–58.90; *p* = 0.078) and previous surgery and infection in the same joint (OR: 10.67; 95% IC: 1.08–105.28; *p* = 0.043). In addition, we observed a higher frequency of recurrence in diabetic patients (OR: 4.87; 95% IC: 0.69–34.50; *p* = 0.113) and those who were treated only with clindamycin (OR: 6.89; 95% IC: 0.80–58.90; *p* = 0.078) than those who only received amoxicillin (OR: 0.357; 95% CI: 0.04–3.55; *p* = 0.379) or rifampin-based combinations (OR: 0.844; 95% CI: 0.12–5.72; *p* = 0.862).

**Table 4.** Comparison of final outcomes.


Regarding surgical treatment, 15/39 patients (38.5%) underwent DAIR, with 13 having favorable outcomes (Figure 2). When analyzed according to both classifications, for patients classified by the Tsukayama guidelines, 9/13 cured patients (69.23%) had type 1 infections and 4/13 (30.77%) had type 2, with one case of recurrence for type 1 and another recurrence for type 2. According to the Zimmerli classification, 12/13 of the cured patients (92.30%) had early infections and 1/13 (7.7%) had a delayed infection, with both recurrences being classified as early infections. Of the 24 patients treated with prosthesis removal, only three had recurrence (12.5%) (Table 4).

**Figure 2.** Flowchart of failure rates according to the medical and surgical approaches used. \* Fifteen patients were treated with a two-stage procedure, six were treated with a one-stage procedure, one was treated with arthrodesis, and two were treated with resection arthroplasty.

Among the 39 evaluable patients, 17 were treated with rifampin. There were no differences in the outcome of patients treated with rifampin-based combinations. There were five patients with positive intraoperative cultures, with one being treated with rifampin therapy and cured, while the others (4/5) did not receive rifampin and recurrences were observed. Of the 15 patients treated with DAIR, eight (53.3%) received rifampin-based regimens, while seven did not, and one recurrence was observed in each group, but this was not significant (Figure 2). We analyzed 11 patients who received clindamycin treatment (six associated with rifampin) and there were three instances of recurrence (all isolates were susceptible to clindamycin).

The epidemiological, clinical, and treatment data of the five patients who showed recurrence are presented in Table 5.


**Table 5.** Individual clinical characteristics and treatment of the five recurrence cases.

Abbreviations: DM, diabetes mellitus; CKD, Chronic kidney disease. <sup>a</sup> Mean age (SD), 60.2 (SD 17.6) years. <sup>b</sup> Tsukayama classification: early postoperative infection (Type 1), late chronic infection (Type 2), and positive intraoperative cultures (Type 4).

#### **3. Discussion**

In this retrospective multicenter study, we described 44 patients with shoulder PJI due to *C. acnes* over a period of 14 years. The diagnosis of this infection is difficult due to the absence of classical clinical evidence, as well as the challenges associated with culturing the microorganism. In this 14 year series, we observed an increase in the number of diagnosed cases of this infection, which is probably due to the extended incubation time that has been demonstrated in other studies for maximizing the recovery of *C. acnes* from PJI specimens [1,7,14–17].

Previous studies have argued that the shoulder has a propensity for infection with *C. acnes* because it is the anaerobic dominant bacteria from healthy skin, particularly in moist areas (axilla), where a higher *C. acnes* bacterial burden is observed in men compared to women [17–19]. Moreover, previous series have reported that male gender is a risk factor for the development of this infection [1,9,20]. These previous findings would explain our results in which a male predominance of PJI was observed.

The most frequent types of infection in this study, according to the Tsukayama classification, were late chronic or positive intraoperative cultures, which is similar to that reported in other studies [1,21,22]; this is due to the paucity of classical symptoms and the absence of elevated inflammatory markers that delay diagnosis. However, when we classified the infection type according to the Zimmerli classification, early infection was the most frequent.

In our study, the most frequent symptom was joint pain. This is consistent with other studies in which pain and functional limitations without either fever or constitutional symptoms were the most frequent clinical presentations [6,7,23].

Previous surgery in the same joint has been linked to an increased risk of *C. acnes*related PJI because repeated manipulation of the joint causes changes in the anatomical structure; this increases the duration of surgery, which is a major risk factor for shoulder PJI from this microorganism [20,24,25]. We observed that previous prosthesis, infection, or surgery in the same joint might be related with recurrence, but we could not demonstrate a significant association, possibly due to small sample size.

In our study, all isolates tested were susceptible to penicillin, vancomycin, and rifampin, with approximately 2.5% being resistant to clindamycin. These observed susceptibility patterns were similar to those of other studies [1,7,22], which suggested that the broad antimicrobial susceptibility of *C. acnes* appeared to be maintained.

Previous clinical studies and case reports provide little information regarding the optimal treatment for *C. acnes*-related PJI. In our study, all patients received antibiotic and surgical treatment. As expected, we observed significant differences between surgery type (DAIR, two-stage surgery, and one-stage surgery), as well as the type of prosthetic infection. Regarding surgical treatment, prosthesis retention and the two-stage procedure were the

most frequent surgical procedures performed, unlike previous articles, which suggests that prosthesis exchange should be the treatment of choice in most cases [2,5,26]. We observed that, in the cases of early infection according to the Zimmerli classification, DAIR treatment may be a safe option.

In terms of antimicrobial treatment, the outcomes with or without adjunctive rifampin therapy were similar to other studies [1,22]. This finding is striking, particularly in cases treated with debridement and implant retention, because this antibiotic has antibiofilm activity and its effectiveness for the eradication of *C. acnes* has been demonstrated both in vitro and in vivo in an animal model of foreign-body infection [12]. However, another explanation could be that the presence of a high inoculum in the biofilms forms in a foreign body (i.e., a prosthetic joint). In this state, the microorganism produces mutations that can lead to some degree of resistance, which is observed as a reduced susceptibility to rifampin; this phenomenon was reported in a study by Furustrand et al. [27], where it was demonstrated in vitro.

On the other hand, we observed a nonsignificant trend toward a higher frequency of failure among 11 patients who received clindamycin (cured 72.7% vs. recurrence 27.3%). The IDSA guidelines recommend clindamycin as an alternative treatment to β-lactams, because the majority of tested isolates were susceptible; for this reason, the use of clindamycin has been evaluated in previous studies [1–3]. However, future clinical trials will be needed to compare antibiotic therapy between β-lactams and clindamycin in *C. acnes*-related PJI.

The strength of this multicenter study is that only patients with a proven diagnosis of *C. acnes-*related PJI were included. Currently, most studies include all types of bone infection due to this microorganism, which makes it difficult to determine the best management and evolution of this entity.

This study did have some limitations. This was a retrospective observational study that did not have predefined therapeutic procedures, and this could have induced bias. Moreover, the follow-up time was limited to a 1 year period. However, in PJIs caused by microorganisms as paucisymptomatic as *C. acnes*, in which DAIR has been performed, a longer follow-up time might be necessary.

#### **4. Methods**

#### *4.1. Study Design, Patients, and Settings*

This multicenter, retrospective observational study was conducted at 16 hospitals belonging to the Prosthetic Joint Infection Group of the Spanish Network for Research in Infectious Diseases between January 2003 and December 2016.

Patients aged 18 years and older with shoulder PJIs that were caused by *C. acnes* and diagnosed between January 2003 and December 2016 were included, regardless of the age of the implant at the time of the initial symptoms. Polymicrobial infections with coagulase-negative staphylococci (CNS) were also included.

#### *4.2. Data Collection*

Cases were identified by searching the databases of previously recorded consecutive PJIs or the general archives at each participating hospital.

Medical chart abstraction was performed using a standardized case report form to retrieve demographic, clinical, and laboratory data. Demographic data included age and sex. Laboratory data included erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and white blood cell (WBC) counts. Clinical data consisted of comorbidities, immunosuppressive therapy, Charlson index, previous exposure to antibiotics (7 days), hospitalization in the previous 90 days (of at least 2 days), and receipt of hemodialysis. We also collected the following information regarding arthroplasty: date of implantation, site, primary or revision arthroplasty, previous infections in the same joint (date and microorganism), cemented versus uncemented arthroplasty, use of antibiotics in bone cement, and date of diagnosis. The time from index surgery to diagnosis was recorded as the time from the last surgical procedure performed pre diagnosis to the first positive *C. acnes* culture, with classification of the PJI, type and number of cultured samples, and their results also being recorded.

Information regarding surgical treatment, exchange of removable pieces of the prosthesis (in at least one debridement surgery), and the type and duration of antimicrobials used was also collected, as well as patient outcomes and the date of the last follow-up visit.

#### *4.3. Definitions*

A PJI was defined on the basis of previously detailed criteria [1,11]. The *C. acnes* etiology was confirmed if ≥2 specimens were positive for *C. acnes*, or if one culture specimen was positive for *C. acnes*, with no other organism detected on culture and concurrent evidence of joint purulence, histopathological inflammation, or a sinus tract communicating with the prosthesis. PJI was assigned according to the Tsukayama and Zimmerli classifications [28–30].

#### *4.4. Follow-Up and Treatment Success*

Antimicrobial therapeutic regimens and treatment outcomes were assessed through the last recorded clinical visit. Decisions on therapeutic regimens were based on the clinical judgment of the infectious disease and surgical specialist providers. The type, delivery method, and duration of antimicrobial therapy were recorded.

After being discharged, patients were followed-up according to the protocol of each participating center. The follow-up period was calculated from surgery due to infection: debridement, one-stage exchange, two-stage exchange, or other procedures (arthrodesis/resection arthroplasty). Among patients in remission, only those with at least 1 year of follow-up were included in the outcome analysis.

Cure was defined as the absence of signs and symptoms of infection at the conclusion of a minimum 1 year follow-up period after antibiotic therapy, which did not result in unplanned additional surgical debridement for putative persistent infection. Treatment failure was established on the basis of the following criteria: (1) persistence of symptoms and clinical signs of infection during treatment that led to a change in the surgical strategy (except for new surgical debridement during the first month after an initial debridement); (2) the recurrence of symptoms and clinical signs of infection once the surgical strategy was completed, with isolation of the same microorganism; (3) the need for suppressive antibiotic treatment against *C. acnes*; (4) infection-related death. Any case of reinfection by microorganisms other than *C. acnes* detected during the follow-up period was not considered a failure.

#### *4.5. Microbiological Methods*

Culture specimens were collected and processed at each participating institution, following the Spanish guidelines for the microbiological diagnosis of bone and joint infections [31,32]. Identification testing of isolates was performed in the clinical microbiology laboratory at each center using standard microbiological techniques. The susceptibilities of *C. acnes* isolates were tested against standard antimicrobial agents. Isolates were classified as susceptible according to the minimum inhibitory concentration (MIC) breakpoints set by the Clinical and Laboratory Standards Institute (CLSI) or the European Committee on Antimicrobial Susceptibility Testing (EUCAST).

#### *4.6. Statistical Analysis*

The descriptive analysis for defining the patient's characteristics was done by frequencies and percentages for categorical variables and measures of central tendency and dispersion for numerical variables. The non-normally distributed continuous variables were expressed by median and interquartile range (IQR). For evaluating the differences between favorable outcome and failed treatment, the Mann–Whitney *U* test was used to compare continuous variables and the chi-squared and Fisher exact tests were used for comparing categorical variables. Moreover, univariate logistic regression was used for

evaluating the recurrence risk. A *p*-value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS software version 26 (IBM Inc., Armonk, NY, USA).

#### **5. Conclusions**

Physicians should be aware of the increase in the frequency of shoulder PJIs caused by *C. acnes* because there are few clinical symptoms and an absence of elevated inflammatory markers. On the other hand, patients with type 1 infections according to the Tsukayama classification or early infection by the Zimmerli classification could be treated with DAIR. According to our data, rifampin therapy does not seem to improve outcomes, and clindamycin seems to be associated with a worse prognosis. Randomized studies with a greater number of patients are necessary to establish the optimal antimicrobial treatment.

**Author Contributions:** Conceptualization and methodology, H.H.V. and M.R.; formal analysis, H.H.V., M.R., and J.L.-T.; investigation, H.H.V.; resources, H.H.V., R.E.-S., M.F.-S., O.M., Á.A., D.R.-P., A.J.-S., M.D.d.T., A.R. (Alicia Rico), L.F., J.P.-S., L.G., J.A.I., J.L.-T., N.B., L.M., A.R. (Antonio Ramirez), M.R., and Antonio Ramirez data curation, H.H.V.; writing—original draft preparation, H.H.V.; writing—review and editing, R.E.-S., M.F.-S., O.M., Á.A., D.R.-P., A.J.-S., M.D.d.T., A.R. (Alicia Rico), L.F., J.P.-S., L.G., J.A.I., J.L.-T., N.B., L.M., A.R. (Antonio Ramirez), and M.R.; visualization, H.H.V., M.R. and J.L.-T.; supervision, M.R. and J.L.-T.; project administration, H.H.V. and M.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work received no specific funding from the public, private, or not-for-profit sectors.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee for Clinical Research of the Balearic Islands, Spain (IB3404/17PI).

**Informed Consent Statement:** Patient consent was waived due to this is a retrospective and observational study, the information is collected from the medical archives of each hospital and there is neither intervention in the patient's treatment nor in the epidemiological and clinical information.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical concerns.

**Acknowledgments:** We acknowledge Aina Millán Pons for the statistical analysis and Antonio Vanrell for the creation of the database. We also thank Jonathan McFarland for improving the English language and to all doctors belonging to the 16 hospitals of the Prosthetic Joint Infection Group of the Spanish Network for Research in Infectious Diseases that participated in this study. GEIO Study Collaborators: Javier Cobo 1, Luis Estelles 2, Julia Laporte 3, Javier Ariza 3, Bernadette G. Pfang 4, Jaime Esteban 4, Carles Pilgrau 5, Mayli Lung 6, Pablo S. Corona 7, Ferrán Perez-Villar 8, Mercé Gracía-Gonzalez 9, Oriol Gasch 10, Maite Ruiz 11, Javier Garcés 12, Mikel Mancheño-Losa 13, Isabel Mur14, Pere Coll 15, Xavier Crusi 16, Alex Soriano 17 1 Infectious Diseases Department, Hospital Universitario Ramón y Cajal, Madrid, España. <sup>2</sup> Infectious Diseases Unit, Department of Medicine, Hospital Universitario Marqués de Valdecilla-IDIVAL, Cantabria, Spain. <sup>3</sup> Infectious Diseases Department, Hospital Universitari de Bellvitge, Barcelona, Spain. <sup>4</sup> Department of Internal Medicine. Bone and Joint Infection Unit. IIS-Fundación Jiménez Díaz. Madrid, Spain. <sup>5</sup> Infectious Diseases Department. Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>6</sup> Microbiology Department. Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>7</sup> Reconstructive and Septic surgery Division. Department of Orthopedic Surgery. Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>8</sup> Orthopedic Surgery and Traumatology Departmet, Hospital Universitari Arnau de Vilanova. Lleida, Spain. <sup>9</sup> Microbiology Department, Hospital Universitari Arnau de Vilanova. Lleida, Spain. <sup>10</sup> Infectious Diseases Department, Corporació Sanitària Parc Taulí, Barcelona, Spain. <sup>11</sup> Microbiology Department, University Hospital Virgen del Rocio, Sevilla, Spain. <sup>12</sup> Orthopedic Surgery and Traumatology Department, University Hospital Virgen del Rocio, Sevilla, Spain. <sup>13</sup> Infectious Diseases Unit, Internal Medicine Department, Hospital Universitario 12 de Octubre, Instituto de Investigación Hospital 12 de Octubre "i + 12", Madrid, Spain. <sup>14</sup> Infectious Diseases Unit. Hospital de la Santa Creu i Sant Pau-Institut d'Investigació Biomèdica Sant Pau; Departament of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>15</sup> Department of clinical Microbiology. Hospital de la Santa Creu i Sant Pau-Institut d'Investigació Biomèdica Sant Pau; Departament of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>16</sup> Department of Orthopedic and Traumatology. Hospital de la Santa Creu i Sant Pau-Institut d'Investigació Biomèdica Sant Pau; Departament of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain. <sup>17</sup> Department of Infectious diseases, Hospital Clínic of Barcelona, IDIBAPS, University of Barcelona, Barcelona, Spain.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Perspective* **Candida Periprosthetic Joint Infection: Is It Curable?**

**Laura Escolà-Vergé 1,2,3,\* , Dolors Rodríguez-Pardo 1,2,3 , Pablo S. Corona 2,3,4 and Carles Pigrau 1,2,3**


**Abstract:** Candida periprosthetic joint infection (CPJI) is a rare and very difficult to treat infection, and high-quality evidence regarding the best management is scarce. *Candida* spp. adhere to medical devices and grow forming biofilms, which contribute to the persistence and relapse of this infection. Typically, CPJI presents as a chronic infection in a patient with multiple previous surgeries and long courses of antibiotic therapy. In a retrospective series of cases, the surgical approach with higher rates of success consists of a two-stage exchange surgery, but the best antifungal treatment and duration of antifungal treatment are still unclear, and the efficacy of using an antifungal agent-loaded cement spacer is still controversial. Until more evidence is available, focusing on prevention and identifying patients at risk of CPJI seems more than reasonable.

**Keywords:** *Candida* spp.; periprosthetic joint infection; fungus; biofilm; antifungal-loaded cement spacer; two-stage exchange surgery

#### **1. Introduction**

Periprosthetic joint infection (PJI), which occurs in approximately 1–2% of all procedures, is one of the most feared complications after arthroplasty due to its associated comorbidities and the possible need for implant removal [1]. Candida periprosthetic joint infection (CPJI) represents a rare etiology among all PJIs; sometimes it is very difficult to diagnose, and it is especially difficult to treat when the prosthetic material cannot be removed [2]. In addition, we have no clear guidelines regarding the best antifungal management in these cases [3–7], and evidence is based on small retrospective series.

#### **2. Epidemiology**

There have been a few recent studies analyzing the prevalence of these infections, and most of them are retrospective in nature [8–11]. A Spanish retrospective multicenter study that analyzed the etiology of PJIs from 2003 to 2012 found that a fungal etiology represented 1.3% of all culture-positive PJIs (*n* = 2288), and *Candida* spp. were responsible for 90% of all fungal infections [9]. A smaller retrospective multicenter study performed in Australia from 2006 to 2008 found that CPJI accounted for 0.7% (1/152) of all culturepositive infections [10], and another study that compared the etiology of PJIs between two referral centers in Europe and in the United States between 2000 and 2011 found that fungal PJIs were responsible for 2.3% of 772 cases and 0.3% of 898 cases, respectively [11].

The species of *Candida* depends on the local epidemiology of the geographical area. In two multicenter studies in Spain [2,9] and one in the United States [12], *C. albicans* was the most frequently isolated fungus (55–65%), followed by *C. parapsilosis* (13–33%).

**Citation:** Escolà-Vergé, L.; Rodríguez-Pardo, D.; Corona, P.S.; Pigrau, C. Candida Periprosthetic Joint Infection: Is It Curable?. *Antibiotics* **2021**, *10*, 458. https:// doi.org/10.3390/antibiotics10040458

Academic Editor: Wolf-Rainer Abraham

Received: 15 March 2021 Accepted: 13 April 2021 Published: 17 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Other species, such as *C. glabrata* and *C. tropicalis,* were more anecdotic (3–7% and 2–4%, respectively). Smaller series have found similar results [13,14], and in a recent review of the literature, *C. albicans* (47.3%) was the most frequent strain isolated, followed by *C. parapsilosis* (22.3%) [15], but epidemiology may still vary among regions.

#### **3. Pathogenesis and Risk Factors**

Colonization by *Candida* spp. is regarded as the first step for subsequent infection [16], and *Candida* spp. are common commensals of the human skin and gut microbiota in healthy individuals [17–19]. Invasive disease, which encompasses both candidemia and deepseated infections, usually results from an abnormal or increased number of fungi combined with alterations in the cutaneous and mucosal barriers due to weakening of host immunity [16,17], which permits the transition from *Candida* sp. commensalism to opportunism. Three possible routes of CPJI development have been described: (1) the hematogenous route from an infected catheter or a urinary or intraabdominal source; (2) direct inoculation during prosthesis implantation, revision surgery, or even after arthrocentesis, especially in colonized patients; and (3) extension into synovial fluid from contiguous infected tissues.

*Candida* spp. have specific properties allowing them to adhere to surfaces and form biofilms, especially on prosthetic devices, which permits the development of persister cells, facilitating antifungal resistance, and explains treatment failure when the implant is not removed. In vitro experiments have shown that *C. albicans* biofilm formation begins with the adherence of yeast to a substrate and thereafter yeast cells proliferate across the surface and produce filamentous forms, including hyphae and pseudohyphae. As the biofilm matures, an extracellular matrix accumulates, facilitating antifungal resistance, notably to azoles and polyenes, through different mechanisms [20], which may explain the high failure rates in CPJI when the implant is not removed. Finally, non-adherent yeast cells are released from the biofilm into the surrounding medium (the dispersal step). *C. albicans*, the most frequent causative agent of CPJI, has been reported to form larger and more complex biofilms than other *Candida* species [21].

All parts of the immune system are involved in the response to this infection. For example, deficiencies in the T-helper 17 lymphocyte cell line impair the mucosal immune response to *Candida* spp. and facilitate Candida infections. Neutrophil dysfunction or leukopenia also predisposes patients to suffer invasive candidiasis, and complement or immunoglobulin deficiency or alteration is associated with complicated disease as well [17]. The regulatory pathways and mechanisms that govern Candida biofilm development are very complex [20]; gene expression of *C. albicans* is regulated by both a continuous host–pathogen interplay and by distinct genetic mechanisms [19], but this is not the scope of this review.

However, there are other factors that are not only easier to identify than alterations in host immunity but also probably more prevalent in patients with CPJIs and may play a major role in the pathogenesis of invasive candidiasis. The most reported factors are as follows [17]: (1) the long-term or repeated use of broad-spectrum antibiotics, especially in the previous 3 months, which depletes commensal gut bacteria, enabling *Candida* sp. overgrowth. Many antibiotics are known to promote fungal growth and pathogenicity because they disrupt the microbiota and eliminate anaerobic bacteria in the gut which could have otherwise inhibited the fungi, and studies show that the introduction of small amounts of *C. albicans* to mice after antibiotic treatment caused significant changes in the gut microbiota, which may persist in the long term [22]. (2) Breach of the cutaneous and gastrointestinal barriers by chemotherapy, surgery, gastrointestinal perforation, or instrumentation, such as central venous catheters, which may facilitate *Candida* sp. translocation into the bloodstream. (3) Immunosuppression secondary to malignant diseases, immunodeficiencies, or immunosuppressive therapy. Other risk factors reported in patients with CPJIs have been older age [18], diabetes, rheumatoid arthritis, malnutrition, and tuberculosis, which probably also reflect alterations in host immunity [2,12–14,23]. Other series have also identified that multiple previous surgeries at the site of the CPJI

are also a risk factor [2,13,23,24]. A recent retrospective case–control study that compared fungal PJIs with bacterial PJIs found that recent antibiotic consumption (OR: 3.4; 95% CI: 1.2–9.3) and prolonged wound drainage (OR: 7.3; 95% CI: 2.02–26.95) were significantly associated with CPJI [13]. In our experience, patients treated with long courses of linezolid for multidrug-resistant chronic bacterial PJIs tend to present mucocutaneous candidiasis, and their colonization may persist for an unknown duration, which could also be another risk factor for hip CPJI.

Although it has not been deeply studied, considering the pathogenesis of the disease, previous *Candida* spp. colonization in the urine or Candida intertrigo may also be risk factors in patients undergoing hip arthroplasty [2,13]. In a multicenter retrospective study of patients with CPJIs, we found 14% of patients with Candida intertrigo and 9% of patients with a previous urinary tract infection (three with positive blood cultures) caused by the same *Candida* spp. before the diagnosis of CPJI [2].

#### **4. Clinical Manifestations and Diagnosis**

CPJIs are usually chronic infections characterized by pain, swelling, and sinus tracts. Implant loosening may be observed on radiography in nearly 50% of cases, as previously reported in some studies [2,25]. In fact, the median duration from the index surgery and the diagnosis of CPJI averaged 17–25 months [12,13]. Blood tests could show no leukocytosis, and the C-reactive protein (CRP) level and erythrocyte sedimentation rate are usually normal or mildly elevated [2,12]. The same recently published study comparing patients with CPJIs with those with bacterial PJIs showed that patients with CPJIs had lower median CRP values (2.95 mg/dL vs. 5.99 mg/dL) and lower synovial fluid leukocyte levels (13,953 cells/mm3 vs. 33,198 cell/mm3) [13].

The criteria to diagnose CPJI are not well established, and the same criteria used in diagnosing bacterial PJIs may not be reliable in some cases. The Infectious Diseases Society of America (IDSA) guidelines [3], a previous International Consensus on PJIs [6], and a recent European Bone and Joint Infection Society (EBJIS) consensus [26] consider two or more intraoperative cultures or the combination of preoperative aspiration and intraoperative cultures yielding the same organism definitive evidence of a PJI [3]. However, when reviewing published series of CPJI cases, the microbiological criteria changed from one study to another. Some authors consider that one positive preoperative aspiration culture and/or a positive intraoperative culture is sufficient [9], while others require two positive cultures [2,13] or one positive culture with additional criteria for PJIs [2,24]. In our opinion, when *Candida* spp. are found in only one intraoperative culture, the case should be evaluated carefully, and treating the Candida etiology should be considered, especially in patients with other risk factors for CPJI such as previous antibiotic therapy or multiple previous surgeries (Figure 1). In fact, even if another microorganism is isolated in two or more cultures, polymicrobial infection is not infrequent, particularly in the hip location, being found in 16% to 26% of cases, depending on the series [2,12], and this should not be a criterion for discarding the value of one positive culture for *Candida* spp.

**Figure 1.** Diagnosis of Candida periprosthetic joint infection.

#### **5. Medical and Surgical Treatment**

International guidelines on candidiasis and PJIs [5,7] recommend, with limited evidence, the combination of prosthesis removal and reimplantation in two stages. They recommend a prolonged period of antifungal therapy for at least 12 weeks after resection arthroplasty and at least 6 weeks after prosthesis implantation, without specifying the best antifungal option [5]. They state that the use of antifungal agent-loaded cement spacers is controversial.

The fact that *Candida* spp. grow and form biofilms on medical devices makes these microorganisms highly resistant to antifungal agents and the host immune system [27–30]. Therefore, the best surgical approach is to remove the prosthetic material to avoid the problem of antifungals penetrating and acting within the biofilm. In this sense, a two-stage exchange arthroplasty strategy is probably the best option when feasible to eradicate the infection and to preserve joint function [15], with variable success rates from 14% to almost 100% depending on the series and on the definition of success [2,12,14,23–25,31–35]. In patients with reduced mobility, particularly old patients with multiple previous surgeries in the same location, a resection arthroplasty may be the best alternative. There is less evidence of success with a one-stage exchange arthroplasty strategy, which has been reported in only a few cases [15,36,37]. In a recent review of the literature of 76 episodes of CPJI, one-stage exchange arthroplasty was performed only in three patients with a favorable outcome [15], but in another series of 11 CPJI episodes, it was performed in four with success in two [14]. However, due to the publication bias, the small amount of experience and the difficulty of curing this type of infection, with a high rate of relapses, in our opinion, this procedure should be used only in very selected cases. Irrigation and debridement with prosthesis retention usually fails to cure the infection (cure rates from 0% to 20%), especially in cases of chronic infection [2,12,23,32,35]. Table 1 summarizes the type of treatment, the duration of follow-up and the outcome of the larger case series (number of patients ≥ 10) of CPJI.

*Antibiotics* **2021**, *10*, 458


andofthe(numberof10)ofCandida

*Antibiotics* **2021**, *10*, 458


138

Fluconazole is active against most CPJI isolates, and it shows good penetration into synovial fluid and less toxicity than amphotericin B, but its activity against *Candida* sp. biofilms is limited. However, the antifungals that have demonstrated better activity against biofilms are echinocandins and liposomal formulations of amphotericin B [27–29,38]. In the absence of clear recommendations for systemic antifungal treatment, the most frequently used antifungals have been fluconazole followed by amphotericin B in older series and [15] by echinocandins in recent series [2], with different outcomes, especially in relation to the type of surgical approach (Table 1). However, due to the rarity of this infection, there will probably not be randomized clinical trials regarding the best antifungal treatment. In our retrospective multicenter study, we found better results when amphotericin B or echinocandins rather than fluconazole were combined with implant removal [2], with remission rates higher than 80% vs. 62%, similar to values reported in previous studies [32,39]. Therefore, we would recommend the use of an antifungal with antibiofilm activity, amphotericin B or an echinocandin, after resection arthroplasty and after prosthesis implantation, following our proposed diagram of treatment in Figure 2.

**Figure 2.** Our proposal for optimal treatment of Candida periprosthetic joint infection.

On the other hand, few studies have evaluated the efficacy of using an antifungal agent-loaded cement spacer in staged exchange arthroplasty for CPJI, so the indication to use it remains controversial. Moreover, there is no consensus on which antifungal agent should be used and at what dose to achieve the optimal balance between cement stability and drug elution. There have been some cases in which amphotericin B deoxycholate or an azole (mainly fluconazole or voriconazole) was mixed with the cement in the spacer [2,15,35,40–42], with different outcomes. In our clinical practice, amphotericin B (200 mg of amphotericin B deoxycholate for every 40 g of bone cement) is often used because of its broad antifungal spectrum and antibiofilm activity, its heat stability, and its availability in powder form. However, amphotericin B has been proven to behave differently than water-soluble antibacterial agents [43,44], and it is not clear whether the local dose is sufficiently high to elute from cement spacers [27,39,42–44] or whether it is toxic to osteoblasts [45]. An in vitro study found that the elution of 800 mg of liposomal amphotericin B was higher than that of the same dose of deoxycholate amphotericin B when mixed with acrylic bone cement, although it was associated with a loss of compressive strength [46]. In addition, some authors and ourselves have concerns about using only antifungal agents in cement spacers, and we prefer to combine amphotericin B with vancomycin plus gentamycin to avoid bacterial superinfections [15]. Until more evidence is available, we believe that using antifungal agent-loaded cement spacers (preferably with amphotericin B and combined with antibacterial agents) in staged exchange arthroplasty seems reasonable to avoid relapses secondary to fungi that may remain adhered to the bone and cement spacer.

Another unsolved issue is the duration of antifungal treatment. Although short antifungal courses (6 weeks) were successful in a small series when using a staged exchange procedure [33], the median duration of antifungal treatment in larger studies was 3 months [2,12,25,34], consistent with IDSA guideline recommendations [5]. In our opinion,

at least 3 months of antifungal treatment are necessary, especially with a drug with antibiofilm activity (an echinocandin or amphotericin B) and combined with implant removal whenever possible, preferably in the form of a two-stage exchange procedure to maintain joint functionality (Figure 2). In patients with high surgical risk for whom prostheses cannot be removed, suppressive therapy with azoles may be an alternative treatment to maintain joint functionality [2].

#### **6. Prognosis and Prevention**

The prognosis of patients with CPJIs varies depending on the medical and surgical approach. Often, aggressive surgical treatment is dismissed due to the patient's comorbidities, and resection arthroplasty or amputation is performed, resulting in poor patient functionality, but at least curing the infection. On the other hand, even if performing the best strategy (a two-stage exchange), some patients may persist with the infection or relapse. A recent study found that the main risk factors for two-stage exchange failure are hemodialysis, obesity, multiple previous procedures, diabetes, corticosteroid therapy, hypoalbuminemia, immunosuppression, rheumatological diseases, coagulation disorders, and infection due to multidrug-resistant bacteria or fungal species [47]. Therefore, if some of these risk factors coexist in a patient with CPJI, a resection arthroplasty, agreed with the patient, may be the best alternative to cure the infection even if it implies loosing functionality. Unfortunately, we have no score of risk that helps us in making the best decision. In addition, due to the formation of biofilms by Candida spp., CPJIs, when treated, may take several months or even years to relapse. Patient follow-up varies among some studies, and this makes it difficult to establish when CPJI can be considered cured. In our personal experience, due to the chronic nature of CPJI, follow-up periods shorter than 2 years may not be able to detect some relapses.

As histories of previous antibiotic therapy or surgery are not modifiable, we believe that searching for and treating Candida intertrigo in patients with risk factors for CPJI would be a reasonable, cost-effective measure [2,13]. Therefore, although there is no strong evidence to support this hypothesis, we believe that patients with previous Candida infection or clinical Candida colonization may benefit from the addition of fluconazole to standard prophylaxis before hip arthroplasty. Another more debatable measure would be including fluconazole in surgical prophylaxis for patients with an advanced age, diabetes, a long course of antibiotic therapy in the previous months (especially if it was with linezolid) and multiple previous orthopedic surgeries. As these factors may be difficult to evaluate retrospectively, prospective multicenter studies are needed.

Given the poor prognosis of this type of infection, until more evidence is available regarding the best antifungal treatment, the duration of treatment, and the efficacy of using antifungal agent-loaded cement spacers, focusing on CPJI prevention remains essential.

**Author Contributions:** L.E.-V. and C.P. contributed to the conception, methodology, writing of the original draft, and writing of the review and editing with the assistance of a medical writer. D.R.-P. and P.S.C. contributed to the review and editing of the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


## *Article* **A New Antifungal-Loaded Sol-Gel Can Prevent** *Candida albicans* **Prosthetic Joint Infection**

**Hugo Garlito-Díaz 1,2, Jaime Esteban 3,\* , Aranzazu Mediero <sup>4</sup> , Rafael Alfredo Carias-Cálix <sup>5</sup> , Beatriz Toirac <sup>6</sup> , Francisca Mulero 7, Víctor Faus-Rodrigo 8, Antonia Jiménez-Morales 6,9 , Emilio Calvo 1,2 and John Jairo Aguilera-Correa 3,\***


**Abstract:** Fungal PJI is one of the most feared complications after arthroplasty. Although a rare finding, its high associated morbidity and mortality makes it an important object of study. The most frequent species causing fungal PJI is *C. albicans*. New technology to treat this type of PJI involves organic–inorganic sol-gels loaded with antifungals, as proposed in this study, in which anidulafungin is associated with organophosphates. This study aimed to evaluate the efficacy of an anidulafungin-loaded organic–inorganic sol-gel in preventing prosthetic joint infection (PJI), caused by *Candida albicans* using an in vivo murine model that evaluates many different variables. Fifty percent (3/6) of mice in the *C. albicans*-infected, non-coated, chemical-polished (CP)-implant group had positive culture and 100% of the animals in the *C. albicans*-infected, anidulafungin-loaded, sol-gel coated (CP + A)-implant group had a negative culture (0/6) (*p* = 0.023). Taking the microbiology and pathology results into account, 54.5% (6/11) of *C. albicans*-infected CP-implant mice were diagnosed with a PJI, whilst only 9.1% (1/11) of *C. albicans*-infected CP + A-implant mice were PJI-positive (*p* = 0.011). No differences were observed between the bone mineral content and bone mineral density of noninfected CP and noninfected CP + A (*p* = 0.835, and *p* = 0.181, respectively). No histological or histochemical differences were found in the tissue area occupied by the implant among CP and CP + A. Only 2 of the 6 behavioural variables evaluated exhibited changes during the study: limping and piloerection. In conclusion, the anidulafungin-loaded sol-gel coating showed an excellent antifungal response in vivo and can prevent PJI due to *C. albicans* in this experimental model.

**Keywords:** sol-gel; anidulafungin; prosthetic joint infection; *Candida albicans*

#### **1. Introduction**

Osteoarthritis is one of the most common musculoskeletal diseases worldwide, and is the most well-known cause of disability among elderly people [1]. The social and economic burden of osteoarthritis-related loss of work is also high [2]. Joint replacement

**Citation:** Garlito-Díaz, H.; Esteban, J.; Mediero, A.; Carias-Cálix, R.A.; Toirac, B.; Mulero, F.; Faus-Rodrigo, V.; Jiménez-Morales, A.; Calvo, E.; Aguilera-Correa, J.J. A New Antifungal-Loaded Sol-Gel Can Prevent *Candida albicans* Prosthetic Joint Infection. *Antibiotics* **2021**, *10*, 711. https://doi.org/10.3390/ antibiotics10060711

Academic Editor: Marc Maresca

Received: 5 April 2021 Accepted: 9 June 2021 Published: 12 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

is a treatment approach that improves quality of life in many individuals worldwide. Although already used routinely, prosthesis implantation is likely to continue to rise in the coming years [3,4]. The primary cause of device failure is prosthetic joint infection (PJI), a disease involving joint prosthesis and nearby tissue. Advances in the study of the transmission, diagnosis, and treatment of PJI over the last 25 years have led to an improvement in outcomes following this difficult complication. PJI occurs rarely (1–2% of all cases), although its effects are often devastating due to the high associated morbidity and substantial costs [2,5,6]. Additionally, the economic burden of PJI is expected to rise in the coming years with increasing life expectancy and the resulting increase in the number of patients undergoing arthroplasty replacements [7].

The most frequently isolated pathogens from PJI are Gram-positive bacteria, especially *Staphylococcus* species, and Gram-negative microorganisms. Nevertheless, other microorganisms, such as fungi, can also cause PJI, particularly *Candida* species [8–13]. *C. albicans* is the most frequent pathogen isolated, followed by *C. parapsilosis* [10]. Most fungal PJIs present with an insidious, chronic clinical course and are associated with risk factors such as advanced age, previous infection with *Candida*, prior antimicrobial use, multiple surgeries on the joint, immunosuppression, and diabetes [10,14–16]. Despite being a rare infection (<1%), up to a quarter of cases can progress to candidemia, which carries an associated mortality of up to 40% [17]. This type of PJI poses a challenge for clinicians and requires a multidisciplinary approach, including systemic antibiotics, local therapies, and surgery [18–20]. The systemic and prophylactic treatment of PJIs may be ineffective, as antimicrobials are incapable of reaching the prosthesis–tissue interface due to the continued presence of necrotic and/or avascular tissue after surgery [21].

To address this problem, local antibiotic therapy was proposed as an alternative and/or adjuvant to systemic prophylaxis or treatment, preventing systemic toxicity and favouring drug release directly within the implant site [22]. Organic–inorganic sol-gels loaded with antifungals were used in this approach. Recently, the incorporation of organophosphate [tris(trimethylsilyl) phosphite] in this sol-gel, made of two silanes (3 methacryloxypropyl trimethoxysilane and 2-tetramethyl orthosilicate), has been shown to enhance the adhesion of sol-gel on metallic surfaces and increase cell proliferation [23]. Recently, some studies have reported the excellent biosafety and bactericidal capacity of these materials, showing that they completely inhibit the formation of biofilm by *S. epidermidis* in venous catheters without deleterious procoagulant effects in the animal model [24]. Furthermore, new studies show that sol-gel coatings loaded with fluconazole can prevent and locally treat yeast PJI, specifically those caused by the *Candida* species [25].

This study aimed to evaluate the efficacy of an anidulafungin-loaded, organic–inorganic sol-gel in preventing PJI caused by *C. albicans* using an in vivo murine model.

#### **2. Results**

#### *2.1. Animal Monitoring*

The median weight of the mice over time by group is shown in Figure 1a,b. Only the group of mice infected with *C. albicans* (Cal 35) after the insertion of an anidulafunginloaded, coated, chemically polished (CP + A) implant showed a significant increase in weight of 44.23 mg per day (*p* = 0.0189). The weight of the remaining groups showed no change over time (*p* > 0.05).

Only two of the six behavioural variables evaluated exhibited changes during the study: limping and piloerection. In the groups with uncoated CP implants, limping decreased significantly over time in both noninfected and Cal35-infected groups (*p* = 0.025, and *p* = 0.026, respectively). The slope of the limping was higher in the Cal35-infected group than in the noninfected one: −0.6694% per day versus −0.5198% per day, respectively (Figure 1c). In both groups of mice with a CP + A implant, limping stayed constant over time (*p* > 0.05) (Figure 1d).

**Figure 1.** Median weight (**a**,**b**), limping (**c**,**d**), piloerection (**e**,**f**), and survival (**g**,**h**) in different noninfected groups (black) and in the Cal 35-infected group (red) with insertion of CP (left column) and CP + A implant (right column) over time.

In the groups of animals with CP implants, only the noninfected group showed a significant decrease in piloerection over time (*p* = 0.031), with a slope of −0.8635% per day (Figure 1e). In the mice with an inserted CP + A-implant, piloerection stayed constant over time in both groups (Figure 1f).

Survival was significantly lower in the Cal35-infected group with CP implants than in the noninfected group as of day 19 (*p* = 0.002) (Figure 1g). Only one mouse (9.1%) in the Cal35-infected group with CP implants died of candidemia (Figure 2). Only one mouse in the Cal35-infected group with CP + A implants died because of a Cytomegalovirus infection (Figure 3); for this reason, this mouse was withdrawn from the survival analysis. Taking this into account, no survival differences were detected between CP + A-implants group and Cal35-infected mice with the CP + A-implants group (Figure 1h).

**Figure 2.** Histological section of the kidney of a mouse belonging to the Cal35-infected CP-implant group that died of candidemia (**a**) and histological sections at higher magnifications in haematoxylin and eosin staining (**b**) and Groccot's stain (**c**,**d**). Black, blue, and red bars represent 2 mm, 50 μm, and 20 μm, respectively.

**Figure 3.** Histological section of the liver of a mouse of the CP Cal35 group that died of CMV infection (**a**) and histological sections with a greater increase in H&E (**b**,**c**) and Groccot's silver stain (**d**,**e**). Black, blue, and red bars represent 2 mm, 50 μm, and 20 μm, respectively. The liver of the animal showed parenchyma with foci of necrosis and a polymorphonuclear-type inflammatory infiltration accompanied by occasional Grocott-positive intracellular inclusions inside some hepatocytes, which was compatible with a cytomegalovirus infection. An acute necrotising inflammatory reaction was detected around the central veins.

#### *2.2. Microbiological and Pathological Results*

The femur culture of the noninfected groups was negative for all of mice. Three of the 11 stamps from Cal35-infected CP-implant mice revealed the presence of yeast in the synovial fluid on Gram staining (Figure 4). All of the stamps from the Cal35-infected mice with CP + A-implants were negative. Each Cal-35-infected group composed of 11 mice was divided into two subgroups: six animals were used for microbiological studies and five for pathological studies.

**Figure 4.** Cytological image of a Gram stain showing a macrophage (**a**,**b**) phagocytizing multiple yeasts (red arrows), and a myocyte (**c**).

Fifty percent (3/6; 95%CI: 0.099–0.900) of mice in the Cal 35-infected group of mice with CP implants had positive culture, whilst 100% of the Cal 35-infected animals with CP + A-implants had a negative culture (0/6) (*p* = 0.023). No statistically significant difference was observed in the quantity of yeast per gram of femur between the 2 Cal35-infected groups (*p* = 0.091) (Figure 5). The mouse that died in the Cal35-infected group with CP implants had granulomas in both kidneys and a concentration of 4.02 log10 (colony-forming units per femur, CFU/femur) on the outside of the operated femur, 5.51 log10(CFU/g) in the operated femur, 5.47 log10(CFU/g) in the kidney, and 25.5 CFU/cm<sup>2</sup> of the implant surface. The renal parenchyma showed extensive Grocott-positive fungal involvement accompanied by intense acute polymorphonuclear-type inflammation, which presented a patchy distribution pattern affecting both the renal cortex and medulla and the pyelocaliceal system (Figure 2).

**Figure 5.** Quantity of yeast per gram of femur from each group of mice.

In the Cal35-infected group with CP implants, acute osteomyelitis was observed in four of the five femurs (Figure 6a); no chronic osteomyelitis was diagnosed, and the presence of

yeast was also detected in four of the five femurs on Grocott's silver staining (Figure 6b). In the Cal35-infected group with CP + A-implants, acute osteomyelitis was observed in two of the five femurs (Figure 6c,d), chronic osteomyelitis was diagnosed in two of the five femurs, and presence of yeast was detected in only one of the five femurs following Grocott´s silver staining; the latter also showed the presence of acute osteomyelitis. Only one mouse (9.1%) in the noninfected group of animals with a CP + A-implant died (Figure 3). The deceased mouse from this group showed signs of having had diarrhoea, enteritis, and hepatomegaly. Furthermore, when the operated femur, both kidneys, and a piece of the liver were sent for microbiological study, no growth in aerobic or facultative anaerobic bacteria or fungi was detected.

**Figure 6.** Histological images with H&E stain (**a**,**c**) and Grocott's stain (**b**,**d**) of Cal-35–infected mice with implant without CP coating (**a**,**b**) and mice infected with implant and anidulafungin coating CP + A (**c**,**d**). The black, green, blue, and red bars represent 2 mm, 200 μm, 50 μm, and 20 μm, respectively.

Taking both the microbiology results and pathology results into account, 54.5% of the Cal35-infected mice with CP implants were diagnosed with a PJI, whilst only 9.1% of the Cal35-infected mice with CP + A-implants were PJI-positive. Therefore, the PJI positivity was significantly higher in the Cal35-infected CP-implant group than in the Cal-35 CP + A-implant group (*p* = 0.011).

The presence of round or ovoid structures accompanied by signs of germination was noteworthy, as was as the presence of other septate structures corresponding to pseudohyphae and hyphae, visible with Grocott's stain. No other infectious agents were observed in the samples studied.

#### *2.3. Microcomputed Tomography and Bone Histology*

No differences were observed between the bone mineral content (BMC) and bone mineral density (BMD) of the groups of mice with CP- and CP + A-implants (*p* = 0.835, and *p* = 0.181, respectively). The BMD results were perfectly comparable as there were no differences in BMC between the compared groups (Figure 7).

**Figure 7.** Bone mineral content (BMC) (**a**) and bone mineral density (BMD) **(b**) and their threedimensional reconstructions of a representative sample of the CP group (**c**) and CP + A group (**d**).

Hematoxilin-eosin staining showed no differences in tissue in the area occupied by the implant among the mice with CP- and CP + A-implants. When bone markers were analysed in the defect area, no changes were observed among the different animals, in alkaline phosphatase (ALP) staining, Cathepsin K or cluster of differentiation 68 (CD68) (Figure 8).

**Figure 8.** Immunohistochemistry for markers of different bone cells. Long bones were processed and immunohistology staining was carried out. Representative images stained for haematoxylin-eosin (H&E), tartrate-resistant acid phosphatase (TRAP) staining, cathepsin K (cath. K), alkaline phosphatase (ALP), and macrophages (cluster of differentiation 68, CD68). H&E images were taken at 4× magnification. All immunostaining images were taken at 10× magnification.

The viable medullary zones are of a habitual trilinear aspect and were arranged in the peripheral ends of the bone (epiphysis).

#### **3. Discussion**

In this study, we demonstrate the in vivo efficacy of anidulafungin-loaded sol-gel to prevent PJI caused by *C. albicans* using a murine model. We describe a novel approach to sol-gel technology applied to Ti materials using anidulafungin to locally prevent the development of yeast biofilms.

The most frequent clinical manifestations of PJI are pain, joint swelling or effusion, erythema or warmth around the joint, fever, drainage, and the presence of a sinus tract connecting to the prosthesis [2]. In our in vivo model, joint pain was evaluated by monitoring mice weight, limping, and piloerection. Weight remained constant over time in all groups, except the Cal35-infected CP + A-implant group, where the mice increased in weight over time. This discrepancy in weight is uncertain, but it could be attributed to the effect of dexamethasone, which has been shown to both increase [26] and reduce mouse weight [27,28]. Moreover, it is known that enrofloxacin does not impede weight gain [29]. Infection seems to decrease limping more significantly, as can be seen in the CP-implant groups. *C. albicans*-caused infections were characterised by a chronic, indolent, and relapsing course [10–12]. This indolent course may be the result of the release of neutrophil extracellular traps (NETs) triggered by farnesol, a crucial quorum-sensing molecule of *C. albicans* [13]. The accumulation of NETs can reduce inflammation through the degeneration of cytokines and chemokines [14]. This could explain why *C. albicans*infected CP-implant mice stopped limping before the noninfected CP-implant animals. In the CP + A-implant groups, limping decreased faster in *C. albicans*-infected mice compared to noninfected mice. This finding is uncertain but could be attributed to the indolent course provoked by the presence of *C. albicans* on the implant, although these yeasts are not viable. Piloerection did not show a clear difference between noninfected and Cal35-infected mice in non-coated or coated implants over time, contrasting with other bacterial PJI in vivo models [30]. The survival of our animal model varied according to the group. One of 11 mice from the Cal35-infected CP-implant group died as a result of renal "fungus balls" caused by *C. albicans* infection [18,19]. These balls are most likely the result of haematogenous seeding from septic arthritis of the knee joint [20]. This highlights the high mortality associated with candidemia derived from *Candida* bone and joint diseases [21,22,31]. However, these results must be interpreted cautiously, particularly when there is a difference of only one individual. Likewise, one of the 11 mice from the noninfected CP + A-implant group perished due to an acute hepato-digestive infection caused by mouse cytomegalovirus. This virus can be latent in different organs (e.g., liver) in immunocompetent mice, and cause acute infection in immunodeficient ones [32]. Furthermore, this virus can be detected as Grocott-positive intranuclear inclusions in pathological samples [33].

The microbiological and histological results obtained in this study revealed the difficulty of inducing this type of infection despite having used two of the most important pharmacological risk factors, i.e., immunosuppression [34] and broad-spectrum antibiotic therapy [35]. This fact could underline the importance of other risk factors, such as systemic disease, diabetes mellitus, revision arthroplasty, type of prosthesis (monoblock or modular), and type of fixation (uncemented, cemented, hybrid, or with plain or antibiotic-loaded cement) [27–29,31,34]. The most important finding of this work is that anidulafungin-loaded sol-gel coating, when applied to orthopaedic implants, can prevent Candida PJI in an in vivo model. Interestingly, some kind of osteomyelitis was detected in three mice, though no presence of yeast was observed. This finding may be due to both the presence of dead yeast killed by anidulafungin and the inhibition of yeast phagocytosis that dexamethasone therapy exerts on phagocytes [36], thereby explaining the inflammation in absence of yeast proliferation. Our results are consistent with other previously published in vitro studies [25]. Moreover, anidulafungin-loaded coating had a non-harmful effect on bone mineralisation according to the microcomputed tomographic images, and no changes in

bone markers were found among groups, thus supporting the results obtained in previous research, based on sol-gel processes [30]. Hence, the fixation of an anidulafungin-loaded sol-gel coated implant is likely to be at least as effective as an uncoated implant.

In recent years, several types of coating have been presented for clinical use: natural, peptide, ceramic, and synthetic coatings [37,38]. Most were designed for osteointegration and antibacterial purposes [30,39–42]. To our knowledge, few studies have developed these coatings to be loaded with antifungals associated with sol-gel technology, as proposed in this work [25]. As fungal PJI prediction is difficult, the use of anidulafungin-loaded sol-gel may be recommended in those patients who have risk factors for developing fungal PJI [35]. This would reduce the personal and healthcare costs associated with this type of infection and its relapses following delayed reimplantation arthroplasty after a follow-up of more than 50 months [43].

However, this study is not exempt from limitations. Firstly, the form of implant infection may have reduced yeast viability on anidulafungin-loaded sol-gel before implantation. No alternative form of infection was possible according to the results obtained in pilot studies (data not shown). Secondly, the results would be more robust with an equal number of samples allocated for microbiological and pathological analyses, although our number of specimens is nearly double that of similar, recently published studies [30]. This unexpected limitation stems from the low infectivity shown by *C. albicans* (approximately 50%), as evidenced in this study. Thirdly, this type of technology can carry other antifungals, e.g., fluconazole [22], which provides it with an antifungal ability against both *C. albicans* and some non-*C. albicans* species, and which should be evaluated using in vivo models for preventive use in some special cases. Fourthly, the death caused by cytomegalovirus suggests that this animal model should be replicated in an Animal Biosafety Level-2 facility, where moderately immunosuppressed animals are less exposed to environmental pathogens.

#### **4. Materials and Methods**

#### *4.1. Sol-Gel Synthesis and Coating of Titanium Implants*

The Ti-6Al-4V implants were made from 0.6-mm thick Kirschner wires provided by Depuy Synthes (Johnson & Johnson, New Brunswick, NJ, USA). Each wire was cut into implants measuring 1 cm in length. Subsequently, these were chemically polished (CP), as previously described [44], to achieve a surface finish more closely resembling that used in routine clinical practice.

Hybrid organic–inorganic sol-gel coatings composed of a mixture of organopolysyloxanes, including methacryloxypropyltrimethoxy silane (MAPTMS, 98%, Acros Organics, Thermo Fisher Scientific, Waltham, MA, USA) and tetramethyl orthosilane (TMOS, 98%, Acros Organics, Thermo Fisher Scientific, Waltham, MA, USA) and biofunctionalised with tris(trimethylsilyl)phosphite (92%, Sigma-Aldrich, St. Louis, MO, USA) were prepared following a previously published methodology [23]. The coating was loaded with 0.99 mg/mL of anidulafungin (Pfizer, New York, NY, USA) by adding the drug to the aqueous phase during its preparation [22]. Finally, the Ti-6Al-4V implants for the in vivo model were coated by dipping them in anidulafungin and allowing them to dry for at least 1 h at 60 ◦C (CP + A).

#### *4.2. Animal Surgical Model and Monitoring*

We used one clinical strain isolated in the clinical microbiology department of Fundación Jiménez Díaz University Hospital: a strain of *C. albicans* from an 81-year-old woman with infection of a hip prosthesis (Cal35). The antifungal susceptibility profile of Cal35 was obtained by the Vitek 2 AST-YS08 yeast susceptibility test (bioMérieux, Marcy l'Etoile, France). Cal35 was susceptible to all antifungals tested, i.e., amphotericin B (≤0.25 μg/mL), caspofungin (≤0.25 μg/mL), flucytosine (≤1 μg/mL), fluconazole (2 μg/mL), micafungin (≤0.06 μg/mL), and voriconazole (≤0.12 μg/mL). Surgical intervention of the in vivo model was based on a modified model previously described by Aguilera-Correa et al. [30]. The intervention consisted of placing the implant into the right femur of RjOrl:SWISS

(CD1) mice (Janvier Labs, France) through the knee using an aseptic surgical technique (Figures 9 and 10). Two main modifications were made. Firstly, all mice were premedicated with 4 mg/L of dexamethasone [45] (B.Braun, Melsungen, Germany) and 0.1 mg/L of enrofloxacin (ganadexil 5%, Industrial Veterinaria, S.A.—Invesa, Spain) [46] in sterile drinking water one week before surgery and for the entire duration of the study. Secondly, the implant infection procedure consisted of incubating 1 mL of a 2.00 McFarland standard of Cal-35 strain in saline (B.Braun, Melsungen, Germany) with each implant in a 12-well plate at 37 ◦C and 5% CO2 for 120 min. After incubation, each implant was rinsed two times in saline. This form of implant infection aims to adhere the yeast to the surface of the implant, due to the impossibility of injecting planktonic yeasts into the femur. In pilot studies conducted prior to this, in vivo model animals infected by planktonic yeasts in the femur before implantation of the prosthesis died from *Candida* infections associated with the liver, kidneys, or lungs (data not shown).

**Figure 9.** Surgical procedure: inhaled anaesthesia and shaving of the limb (**a**), antiseptic washing and isolation of the surgical field (**b**,**c**), skin dissection and exposure of the bony entry point (**d**), retrograde introduction of the biomaterial into the femur of the mouse (**e**), suture and cleaning of the surgical wound (**f**,**g**), awakening and care of the animal (**h**).

Sixteen-week-old male mice with femoral implants were randomly distributed into four groups: one group with a CP implant without infection (CP group, *n* = 11), a group with a CP implant with infection induced by Cal35 (CP Cal35 group, *n* = 11), the third with a CP implant coated with anidulafungin-loaded sol-gel without infection (CP + A group, *n* = 11), and the fourth with a CP implant coated with anidulafungin-loaded sol-gel with infection induced by Cal35 (CP + A Cal35 group, *n* = 11). The sample size was estimated by Wilcoxon Mann–Whitney test and an a priori type of power analysis, considering *d* = 1.5, α = 0.05, (1-β) = 0.95, allocation ratio = 1 by using G\*Power 3.1.9.7 software [47]. The *d* parameter is based on the assumption that the anidulafungin-loaded coating is able to reduce the yeast concentration by at least 80% per gram of bone when compared to the uncoated implant group. The statistical power of the sample was 0.9522. All the animals were included in the study and there were no exclusions.

**Figure 10.** Fluoroscopy of a mouse with an implant placed in the femur.

We assessed the pain-stress and weight of each animal every 48 h on weekdays to ensure physical status. Evaluation of pain/stress was based on the presence or absence of six directly related behaviours in this species for the surgical procedure the animals underwent, i.e., limping, piloerection, lack of grooming, wound presence, passivity, and aggressiveness. In cases of sustained weight loss over time, the most appropriate refinement measures were taken to encourage the animal to eat. For this, they were offered an additional mixture of grains and vegetables (Vitakraft, Bremen, Germany). Five weeks (35 days) after surgery, all the animals were euthanised using hypercapnia. The right femur of each animal was then removed following sterile preparation of the knee, and the samples were sent for analysis. In case of the pre-euthanasia death of one of the mice in any group, the operated femur was alternatively used for microbiological or pathological studies.

#### *4.3. Microbiological and Pathological Studies*

After euthanasia and previous extraction of the femurs from Cal35-infected mice, joint fluid samples were taken using sterile swabs, and this fluid was used to make stamps on a slide for Gram staining. The 11 mice in the Cal-35-infected group were divided into two subgroups: 6 animals were used for microbiological studies and 5 for pathological studies. Three femurs from each noninfected group were used for microbiological studies.

Extracted bones were processed according to the methodology previously described by Aguilera-Correa et al. [30]. Briefly, using a hammer, each femur was divided into two samples in a sterile bag: (1) bone and adnexa and (2) implant. The bone was immersed in 2 mL of saline and sonicated using a sonicator at 22 ◦C for 5 min [48]. The resulting sonicate was diluted in a 10-fold dilution bank and seeded on chloramphenicol-gentamicin Sabouraud agar (bioMérieux, Marcy l'Etoile, France) using the plaque extension method, which consists of seeding 100 μL/plate of each dilution. The concentration of yeasts was estimated as CFU/g of bone and adnexa. The implant was sonicated in 2 mL of saline for 5 min to release the adhered yeast biofilm and to estimate biofilm concentration, measured as CFU/cm<sup>2</sup> of the implant. All plates were checked at 48 and 72 h.

The five femurs obtained from the Cal35-infected group were fixed in 4% paraformaldehyde for 48 h, decalcified in 10% ethylenediaminetetraacetic acid (EDTA) for 4 weeks, paraffin-infiltrated, and stained with haematoxylin-eosin. The presence of some necrotic trabecular bone and some repair areas, fibrosis, and adipose replacement of the bone marrow were identified and recorded. The presence of yeast was determined by using

Grocott's silver stain [49]. The presence/absence of round or ovoid structures, with or without signs of germination, was recorded.

Histopathological definitions were as follows:


#### *4.4. Microcomputed Tomography*

Eight bone samples from each noninfected group included in the aforementioned model were fixed in 10% formaldehyde for 48 h at 4 ◦C. After fixation, the samples were dehydrated in 96% ethanol for 48 h, changing the ethanol every 24 h, and in 100% ethanol for 48 h, changing the ethanol every 24 h. Hind legs were removed and fixed in 10% neutral buffered formalin. Before CT scanning, the paws were washed with running water for 15 min. Three-dimensional microcomputed tomographic imaging was performed with a CompaCT scanner (SEDECAL Madrid, Spain). Data were acquired with 720 projections by 360◦ scan, the integration time of 100 ms with three frames, a photon energy of 50 KeV, and current of 100 uA. The duration of imaging time was 20 min per scan. Three-dimensional renderings of images of hind paws were generated through original volumetric reconstructed images by MicroView software (GE Healthcare, Boston, MA, USA). Comparable regions of interest consisting of three metatarsal joints from each mouse were selected for analysis. Bone volume (BV), bone mass (BM), BMD (calculated as BM/BV mg/cm3), and mean cortical thickness (mm) were quantified from micro-CT scans using GE MicroView software v2.2.

#### *4.5. Immunohistochemistry*

Five out of eight femurs used for microcomputed tomography from each group were decalcified in 10% EDTA for 4 weeks, paraffin-infiltrated, and stained with haematoxylineosin. In the noninfected groups, implants were removed and transversal sections in the knee condyles (5 μm) were made. Immunohistochemical analysis was carried out as previously described [50]. Briefly, sections were incubated with proteinase K Solution (20 μg/mL in Tris-EDTA Buffer, pH 8.0) for 15 min in a water bath at 37 ◦C for antigen retrieval after deparaffinisation and re-hydration. The blocking of nonspecific binding was performed with phosphate buffer saline (PBS), 3% bovine serum albumin (BSA) and 0.1% Triton X-100 for 1 h, and the primary antibodies anti-cathepsin K (1:25), cluster of differentiation 68 (CD68) (1:200), and alkaline phosphatase (ALP) 1:200 (all antibodies from Santa Cruz Biotechnology, Santa Cruz, CA, USA) were incubated overnight at 4 ◦C in a humidifying chamber. The secondary antibodies goat anti-rabbit-fluorescein isothiocyanate (FITC) (1:200) and goat anti-mouse FITC (1:200) (Invitrogen, Life Technologies, Carlsbad, CA, USA) were incubated for 1 h in the dark. Slides were mounted with Fluoroshield with 4 ,6-diamidino-2-fenilindol (DAPI) mounting media (Sigma-Aldrich, St. Louis, MO, USA). Images were taken with the iScan Coreo Au scanner (Ventana Medical Systems, Roche Diagnostics, Basel, Switzerland) and visualised with Image Viewer v.3.1 software (Ventana Medical Systems, Roche Diagnostics, Basel, Switzerland). Images were taken at 4× or 10× magnification.

#### *4.6. Statistical Analysis*

The primary hypotheses were such that the anidulafungin-loaded sol-gel would prevent *C. albicans* PJI and that the anidulafungin-loaded sol-gel can be used without altering bone metabolism at any level. The secondary hypothesis was that the effect of anidulafungin-loaded sol-gel in preventing PJI could be evaluated by animal monitoring before euthanasia.

Statistical analyses were performed using Stata Statistical Software, Release 11 (Stata-Corp, College Station, TX, USA). Data were evaluated using a one-sided Wilcoxon nonparametric test to compare 2 groups or a one-sided proportion comparison Z-test. A log-rank test was used to perform a pairwise comparison of the Kaplan–Meier survival curves of two groups. Statistical significance was set at *p* ≤ 0.05. Body weight was evaluated over time using a linear regression model. Microbiological results and weight values are represented as median and interquartile range. Other behavioural variables are represented as relative frequencies at each time point.

#### **5. Conclusions**

In conclusion, anidulafungin-loaded sol-gel can prevent PJI caused by *C. albicans* without compromising bone integrity.

#### **6. Patents**

The sol-gel used in this study is one of the products protected by the Spanish patent system with Publication Number 2686890, applied for 19 April 2017, and entitled Procedure for Obtaining a Sol-Gel Coating, Composition Coating and Use of the Same.

**Author Contributions:** Conceptualisation, J.J.A.-C. and J.E.; methodology, J.J.A.-C., H.G.-D. and A.M.; software, J.J.A.-C., A.M., R.A.C.-C. and F.M.; validation, J.J.A.-C., J.E., F.M. and R.A.C.-C.; formal analysis, J.J.A.-C., H.G.-D. and R.A.C.-C.; investigation, J.J.A.-C., H.G.-D., A.M., R.A.C.-C., B.T., F.M., V.F.-R., A.J.-M., E.C. and J.E.; resources, J.J.A.-C., H.G.-D., A.M., R.A.C.-C., B.T., F.M., V.F.-R., A.J.-M., E.C. and J.E.; data curation, J.J.A.-C., H.G.-D., A.M., R.A.C.-C., F.M. and J.E.; writing original draft preparation, H.G.-D., J.J.A.-C., A.M., F.M. and R.A.C.-C.; writing—review and editing, J.J.A.-C., H.G.-D., A.M., R.A.C.-C., B.T., F.M., A.J.-M., E.C. and J.E.; visualisation, J.J.A.-C. and H.G.-D.; supervision, A.J.-M., E.C. and J.E.; project administration, J.J.A.-C., A.J.-M. and J.E.; funding acquisition, J.E., A.J.-M. and J.J.A.-C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received financial support from the Mutua Madrileña Foundation (04078/001).

**Institutional Review Board Statement:** The use of yeast strains does not require approval by any research committee according to legislation in force. This study was approved by the Instituto de Investigación Sanitaria Fundación Jiménez Díaz (IIS-FJD) Animal Care and Use Committee (IRB number 04078-001), which includes ad hoc members for ethical issues. Animal care and maintenance complied with institutional guidelines as defined in national and international laws and policies (Spanish Royal Decree 53/2013, authorisation reference PROEX019/18 8 March 2018 granted by the Counsel for the Environment, Local Administration and Territorial Planning of the Community of Madrid and, Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data supporting reported results can be found by contacting with the corresponding authors.

**Acknowledgments:** We wish to acknowledge. Oliver Shaw for reviewing the manuscript for language-related aspects. The authors are also grateful to the Experimental Surgery and Animal Research Service, specifically Carlos Castilla-Reparaz and Carlos Carnero-Guerrero.

**Conflicts of Interest:** Jaime Esteban received travel grants from Pfizer and conference fees from bioMérieux and Heraeus. The funders had no role in the design of the study, the collection, analyses, or interpretation of data; the writing of the manuscript, or in the decision to publish the results. The remaining authors declare no conflict of interest.

#### **References**


*Article*
