https://orcid.org/0000-0001-5232-7900
Figures
Abstract
Context
Candida-related infections are nowadays a serious Public Health Problem emerging multidrug-resistant strains. Candida biofilm also leads bloodstream infections to invasive systemic infections.
Objective
The present meta-analysis aimed to analyze Candida biofilm rate, type, and antifungal resistance among hospitalized patients between 1995 and 2020.
Data sources
Web of Science, Scopus, PubMed, and Google Scholar databases were searched for English papers using the following medical subject heading terms (MESH): “invasive candidiasis”; “bloodstream infections”; “biofilm formation”; “biofilm-related infections”; “mortality”; and “prevalence”.
Study selection
The major inclusion criteria included reporting the rate of biofilm formation and the prevalence of biofilm-related to Candida species, including observational studies (more exactly, cohort, retrospective, and case-control studies). Furthermore, data regarding the mortality rate, the geographical location of the study set, and the use of anti-fungal agents in clinical isolates were also extracted from the studies.
Data extraction
Independent extraction of articles by 2 authors using predefined data fields, including study quality indicators.
Data synthesis
A total of 31 studies from publicly available databases met our inclusion criteria. The biofilm formation in the data set varied greatly from 16 to 100% in blood samples. Most of the studies belonged to Europe (17/31) and Asia (9/31). Forest plot showed a pooled rate of biofilm formation of 80.0% (CI: 67–90), with high heterogeneity (Q = 2567.45, I2 = 98.83, τ2 = 0.150) in random effects model (p < 0.001). The funnel plot and Egger’s linear regression test failed to find publication bias (p = 0.896). The mortality rate in Candida-related bloodstream infections was 37.9% of which 70.0% were from biofilm-associated infections. Furthermore, Candida isolates were also characterized in low, intermediate, or high biofilm formers through their level of biofilm mass (crystal violet staining or XTT assays) after a 24h growth. When comparing between countries, statistical differences were obtained (p = 0.0074), showing the lower and higher biofilm prevalence values in Italy and Spain, respectively. The prevalence of low, intermediate, and high biofilms were 36.2, 18.9, and 35.0% (p < 0.0001), respectively. C. tropicalis was the prevalent species in high biofilm formation (67.5%) showing statistically significant differences when compared to other Candida species, except for C. krusei and C. glabrata. Finally, the rates of antifungal resistance to fluconazole, voriconazole, and caspofungin related to biofilm were 70.5, 67.9 and 72.8% (p < 0.001), respectively.
Citation: Atiencia-Carrera MB, Cabezas-Mera FS, Tejera E, Machado A (2022) Prevalence of biofilms in Candida spp. bloodstream infections: A meta-analysis. PLoS ONE 17(2): e0263522. https://doi.org/10.1371/journal.pone.0263522
Editor: Surasak Saokaew, University of Phayao, THAILAND
Received: July 28, 2021; Accepted: January 20, 2022; Published: February 3, 2022
Copyright: © 2022 Atiencia-Carrera et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by COCIBA Research Grant 2018-2019 through project ID: 12260 entitled “Adhesión inicial y resistencia antimicrobiana de Candida sp. aisladas de la microbiota humana”, under regulations of the Ministry of Health of Ecuador (Contrato Marco de Acceso a los Recursos Genéticos No. MAE-DNB-CM-2016-0046).
Competing interests: The authors have declared that no competing interests exist.
Introduction
Invasive candidiasis is a systemic mycosis caused by Candida species, being commonly described as an opportunistic infection. The population group more vulnerable for invasive candidiasis includes patients with critical illness or immunosuppression (such as hematological and solid organ malignancy, hematopoietic cell and solid organ transplantation, recent abdominal surgery, and hemodialysis), or even people with a central venous catheter, parenteral nutrition. In addition, people that received broad-spectrum antibiotics or with drug habits are also susceptible to invasive candidiasis, as well as premature newborns [1]. All these plausible scenarios lead this systemic infection to be nowadays the 4th leading nosocomial infection in the United States, demonstrating mortality of up to 40% [2]. In Europe, Bassetti and colleagues realized a multinational and multicenter study in 2019 reporting 7.07 episodes per 1000 in European intensive care units (ICUs) with a 30-day mortality of 42% [3]. While, in the Asia-Pacific region, Hsueh and colleagues reported a candidemia incidence in ICUs of 5- to 10-fold higher than in the entire hospital and a mortality rate of patients between 35% and 60% [4]. In Latin America, Nucci and colleagues realized a laboratory-based survey between November 2008 and October 2010 among 20 tertiary care hospitals in seven Latin American countries, reporting an overall incidence of 1.18 cases per 1,000 in general admissions [5]. The mortality associated with invasive candidiasis is similar or even higher in other worldwide countries [6].
To understand the dimension of this infection and its virulence, we must define the term invasive candidiasis as both forms of candidemia detected in the blood and tissues or deep organs under the mucosal surfaces (also known as deep candidiasis). Deep candidiasis can remain localized or spread causing a secondary infection [7]. Patients with a systemic infection induced by Candida spp. can be subdivided into three groups: (1) those who present with bloodstream infection (candidemia); (2) those who develop deep-seated candidiasis (most frequently intra-abdominal candidiasis); and, (3) those who develop a combination of these two groups [8].
The gold standard for the diagnosis of invasive candidiasis is the growth culture, being blood culture commonly used to diagnose candidemia while culture media is applied to diagnose deep candidiasis from tissue biopsies [9]. In this meta-analysis, we only evaluated studies using positive blood cultures to evaluate the biofilm formation and other related factors in candidiasis virulence. More exactly, the selected studies performed an in vitro biofilm assay using Candida isolates from blood samples of the patients with catheter-related candidemia (CRC) and non-CRC. In cases of patients with CRC, the standard procedure was blood cultures from obtained the catheter and peripheral veins, whereas non-CRC was indicated by the recovery of Candida spp. from only blood samples, as previously described by Guembe and colleagues [10].
Nosocomial infections are closely associated with biofilms growing attached to medical devices or host tissues [11]. Biofilms are the predominant growth state of many microorganisms, being a community of irreversible adherent cells with different phenotypic and structural properties when compared to free-floating (planktonic) cells. National Institutes of Health estimated that biofilms are responsible, in one way or another, for more than 80% of all microbial infections in the United States [12]. Candida species can produce well-structured biofilms composed of multiple types of cell and even microbial species, leading to an intrinsic resistance against a wide variety of stress factors, such as various antifungal drugs and immune defense mechanisms [13]. Although the dynamics biofilm-host is not yet fully understood, it is well-known that Candida biofilms inhibit the innate immune system of the host [14]. Therefore, our main goal was to analyze the relationship between biofilms and mortality in Candida spp. related infections, showing a severe menace to the Public Health System with serious outcomes.
Results
Study inclusion criteria and characteristics of the eligible studies
A total of 214 studies were retrieved and 70 full texts were reviewed from publicly available databases (Web of Science, Scopus, PubMed, and Google Scholar). Thirty-one studies met our inclusion criteria (Fig 1). The final data set included studies covering different global regions (most of them in Europe). All available and relevant data were extracted of each study, more exactly, biofilm rate, biofilm type, underlying disease of the patients, Candida species reported, and antifungal resistance. The data was then used to create other databases, collecting information of at least five or more papers, and consequently, each paper was cited more than once. These additional databases were chosen to realize subgroup analysis using a random-effect model and to answer relevant questions about Candida-related biofilms, such as the mortality rate related to biofilms, the geographical distribution of biofilms, the characterization of biofilm production among Candida species, and the correlation between biofilm formation and antifungal resistance (S1 and S2 Files).
As shown in Fig 1, a total data set of 31 studies was achieved for the present meta-analysis following the eligibility criteria, screening process, and quality assessment.
Overall effects of Candida biofilms
The data set reported biofilm rates of Candida-related infections among hospitalized patients between 1995 and 2020 in several countries worldwide. As shown in Table 1, the biofilm formation by Candida spp. isolates in the data set varied greatly from 16% to 100% in blood samples from hospitalized patients. Most of the data set belonged to studies realized in Europe (17/31), followed by Asia (9/31), South America (3/31), and North America (2/31).
Although the methodologies to quantify biofilm biomass varied between studies, these methodologies are based on the optical density (OD) obtained by the combination of a certain colorimetric compound or a simple dissolution in a buffer or water with the growth of the isolated Candida sp. and then it’s compared with reference Candida strains in the same growth conditions. The main methodologies in our study set were crystal violet (CV) assays using microplate reader (51.6%; 16/31), assays with tetrazolium dye (2,3-bis-(2-methoxy-4-nitro-5-sulphenyl)-(2H)-tetrazolium-5-carboxanilide, XTT) using micro plate reader (35.5%; 11/31), and Branchini’s method (9.7%; 3/31). The Branchini’s method, also called slime production method, is based on the production of a viscid slime layer by the growth of the Candida isolate in a tube containing Sabouraud broth [15].
Regardless of the applied methodology in the studies, all these authors were able to evaluate biofilm formation among Candida isolates. However, only 18 of 31 studies were able to categorize the biofilm formation, and so just 5 studies were able to evaluate a positive correlation between biofilm presence and increment of antifungal resistance in the treatment. Finally, the incidence of mortality among patients varied considerably among studies, reporting the values of attributable mortality between 6.9 and 70%. All the information extracted is available in the supplementary section.
Analysis of the forest plot was then realized with data set, showing a pooled rate of biofilm formation of 80.0% (CI: 67–90), as shown in Fig 2. The heterogeneity indices obtained using random effects model (p < 0.001) were Q = 2567.45 (p < 0.001), I2 = 98.83, and τ 2 = 0.150. The pooled rate of biofilm formation obtained needs to be carefully analyzed given the high value of heterogeneity. This will be addressed in our discussion.
A funnel plot was realized to evaluate the existence of publication bias in the final data set (Fig 3). Furthermore, Egger’s linear regression test was also used to reveal any publication bias and possible asymmetric data distribution in the funnel plot.
Studies are represented by a point. The X-axis represents the effect size (biofilm prevalence), and the Y-axis shows the standard error. Despite some asymmetry revealed by the funnel plot in the data set, Egger’s test failed to show publication bias (p = 0.896).
No publication bias was identified by the Egger’s linear regression test (p = 0.896). However, as we will discuss in the next section the qualitative analysis of the funnel clearly suggests some biases from the departure of the geometry from the expected triangular form. The funnel plot of this study illustrates the effect size (biofilm prevalence) on the x-axis and the standard error (SE) on the y-axis. In case of no publication bias in the data set, the studies are distributed evenly around the pooled effect size. The smaller studies should appear near the bottom due to their higher variance when compared to the larger studies, which should be placed at the top of the plot. The diagonal lines show the expected 95% confidence intervals around the summary estimate. In the absence of heterogeneity, the studies of the data set should lie within the funnel defined by these diagonal lines. However, heterogeneity and some asymmetries among the studies of the data set were illustrated by the funnel plot. In our case, we found studies with low errors (similar sizes) but with drastic differences in the biofilm prevalence. This type of pattern probably indicates the presence of confounding variables (sub-groups undelaying structures) which are not included in the global analysis.
Although an obvious biofilm prevalence was found in the data set, the selected studies poorly described the underlying conditions of the patient with biofilm production. The analysis of these conditions among the patients was merely descriptive, as shown in Table 2.
The lack of a detail description of the clinical background and host factors in the patients among the studies represents a main drawback of the present meta-analysis precluding the evaluation of clinical or patient factors and the ability of Candida isolates to establish biofilm. Nonetheless, the ability to establish biofilm is a virulence factor by itself and should be evaluate as risk factor in the treatment of patients with Candida-related blood infections. As summarized in Table 2, only 16 of 31 studies reported some sort of clinical background of the patients with Candida-related bloodstream infections. From this subset of studies, patients evidenced mainly the following clinical conditions: hematological or solid cancer (68.8%, 11/16), surgery interventions (62.5%, 10/16); patients with central venous catheter (56.3%, 9/16); adults under total parenteral nutrition (50.0%, 8/16); patients with human immunodeficiency virus (HIV; 50.0%, and 8/16); patients with diabetes (43.8%; 7/16); patients in the intensive care unit (ICU; 37.5%, and 6/16); patients with immunosuppressive therapy (37.5%, 6/16) and, the remaining clinical backgrounds were only described in 25% or less of the studies in this subset, such as neutropenia (4/16), cardiovascular diseases (3/16), pulmonary diseases (3/16), urinary catheter (3/16), chemotherapy (2/16), and renal insufficiency (2/16). The heterogeneity of the clinical background of the patients and the gap of the host epidemiological factors in these studies excluded further analysis between Candida-related biofilm isolates and clinical history.
Mortality among patients with Candida biofilm
Further subgroup analysis using a random-effect model was realized to differentiate the Candida-related mortality rates between bloodstream infections with planktonic cells and biofilm formation. From the initial data set, only 15 studies evaluated the mortality among patients with Candida-related bloodstream infections. As shown in Table 3, the pooled mortality rate due to Candida-related bloodstream infections was 37.9% (95% CI: 26.2–50.2) of which the mortality associated with biofilm-forming infections was 70.0% (95% CI: 52.8–84.8).
In both scenarios, the mortality rate was statistically incremented among hospitalized patients (p < 0.0001). However, biofilm-related infections evidenced almost the double value of mortality rate in patients, when compared to all Candida-related bloodstream infections.
Geographical distribution of biofilm-forming Candida spp. isolates
The prevalence rate of biofilm-related infections significantly varied among studies of different countries and regions. Therefore, a subgroup analysis was realized between the biofilm formation rates and the geographical region to evaluate possible statistically significant differences (Table 4). Subgroup analysis evaluated the biofilm prevalence between regions and countries with a minimum of published studies, at least two and three studies per region and country, respectively. However, Egger’s test was not applied due to the low number of studies in this analysis.
Although the biofilm prevalence varied among regions, there were no statistically significant differences (p = 0.4049). Europe reported a greater number of studies and showed an intermediate biofilm prevalence among Candida spp. infections. Meanwhile, when comparing prevalence rates between countries, a statistically significant value was obtained (p = 0.0074). In the pairwise comparison analyses, Spain was significantly superior to Brazil (p < 0.0001), Italy (p = 0.0263), and India (p = 0.0030).
Biofilm-forming capability in Candida spp. isolates
Candida spp. isolates vary in their ability to form biofilms, being usually categorized as low (LBF), intermediate (IBF), and high biofilm formers (HBF) according to biomass production (S1–S3 Figs). Briefly, biofilm forming capacity was assessed using the crystal violet or XTT assays, measuring the biofilm mass. Candida isolates were cultured in 96-well plates at 37°C for 24 h and the biomass of each isolate was measured. Then, isolates were grouped based on their level of biomass, more exactly: low biofilm formers (LBF) showed a biomass production below the 1st quartile (Q1; Absisolate < 0.432), intermediate biofilm formers (IBF) evidenced a biomass production in the 2nd quartile (Q2; 0.432 < Absisolate < 1.07), and high biofilm formers (HBF) demonstrated a biomass production higher the 1st quartile 3rd quartile (Q3; Absisolate > 1.07), as previously described by Monfredini et al. [16] and Vitális et al. [17]. Eighteen studies reported this biofilm classification and so a subgroup analysis was realized (Table 5).
Statistically significant differences were found among Candida isolates according to their biofilm-forming capability (p < 0.0001), evidencing a low number of Candida isolates related to intermediate biofilms. No publication bias was detected in both subgroups according to Egger’s linear regression test.
Evaluation of biofilm formation between different Candida species
Although Candida spp. isolates vary in their ability to form biofilms, little is known about this biofilm-forming ability among Candida species. Each category of biofilm was further evaluated among Candida species to evaluate the most virulent Candida species (S1 Table). When analyzing HBF (Table 6), C. tropicalis was the most prevalent HBF overpassing C. albicans and C. parapsilosis by a factor of 2. More precisely, the HBF prevalence of C. tropicalis was the highest showing statistically significant differences with the other Candida species, except for C. krusei (p = 0.5477) and C. glabrata (p = 0.0896).
In order to comprehend how these two major factors: countries and Candida species could actually explain the high heterogeneity showed in our data, we carried out a meta-regression analysis. The inclusion of both variables as interacting variables in a multiplicative model (R2 = 59.13%, p < 0.0001) explained more than an additive model (R2 = 43.48%, p < 0.0001), regarding the prevalence of biofilm formation.
Evaluation of antifungal resistance pattern among Candida isolates
Multiple antifungal resistance among candidiasis has become a serious public health issue, leading to clinical complications and expensive costs. A subgroup analysis based on antifungal resistance was also realized among our study set. Due to the different methodologies used to test susceptibility, the number of studies not enough to analyze statistically antifungal resistance rates between Candida species. As shown in Table 7, the rates of antifungal resistance to fluconazole, voriconazole, and caspofungin related to biofilm-forming strains were 70.5, 67.9, and 72.8%, respectively.
When comparing to planktonic cells, all Candida-related biofilm isolates showed a statistical increment of resistance against the three antifungals evaluated in the study (p < 0.001).
Discussion
The present study evaluated a possible relationship between Candida-related biofilm formation, bloodstream infections, and mortality among hospitalized patients. Invasive mycoses are responsible every year for more than two million infections worldwide and for, at least, as many deaths as tuberculosis or malaria. Candidiasis, aspergillosis, cryptococcosis, and pneumocystosis cause more than 90% of reported deaths associated with invasive mycoses [18]. Among them, the most frequent mycosis is invasive candidiasis causing high morbidity in critically ill patients [19].
Overall effects of Candida biofilms in infections and mortality
As previously referred, around 70.0% of candidemia reports were caused by biofilm-forming strains. However, its biofilm formation was less than in isolates from urogenital infections [20–23] and even respiratory tract infections [22, 23]. Still, the rate of candidemia-associated biofilm infections was higher than oral-related biofilm infections [24] and more than invasive infections [25]. These findings are in agreement with the Institute of Health in the United States, which estimates that biofilms are responsible, in one way or another, for over 80% of all microbial infections [12]. Yet, the reports of Candida-associated biofilm infections varied greatly between published studies possibly due to the lack of differentiation between Candida species, the experience of the researchers, the number of Candida isolates in the study set, and the diversity of biofilm detection and quantification methodologies and its subsequent classification within the study set, such as crystal violet assay, biomass measure, XTT reduction assay, and microtiter plate method [8, 12].
Another issue concerns the lack of differentiation between planktonic and biofilm-related Candida infections in the diagnosis of the clinical laboratories at public health system [19, 26]. The traditional clinical microbiology laboratories have focused on testing planktonically isolated microorganisms and reporting the susceptibility to various antimicrobials under planktonic growth conditions [27]. While the authors from the studies of this meta-analysis applied a further analysis by evaluating the ability of biofilm production in Candida isolates through an in vitro biofilm assay. In Candida biofilms, traditional techniques require device removal followed by culture or microscopy of a catheter segment, while catheter-sparing diagnostic tests include paired quantitative blood cultures. However, as previously indicated by Høiby et al. (2015) and Bouza et al. (2013), the number of positive peripheral blood cultures also seems to be a promising diagnostic tool to diagnose catheter-related candidemia without directly removing the catheter [27, 28]. Therefore, an implementation of a new gold standard methodology is vital to a better characterization of microbial-associated infections avoiding unproductive treatments among hospitalized patients. The mortality rate caused by biofilm formation in Candida-related infections was almost double when compared to planktonic infections. Other studies already stated the burden of invasive candidiasis and its severe outcomes [1, 29], indicating biofilm formation and antifungal resistance as main risk factors among patients. Moreover, we report a pooled attributable mortality of 37.9% to Candida-related bloodstream infection with planktonic cells, which is in agreement with previous reports [1, 18, 30, 31]. These studies reported a mortality range between 25 and 40%, showing a higher mortality incidence among ICU or burn patients, and immunocompromised patients [32]. While the mortality associated with biofilm-forming strains was 70.0% in Candida-related bloodstream infections. However, this correlation has been debated by several authors [10, 16, 33, 34], reporting different mortality rates (25–70%).
It is also important to mention that the ability to quicky proliferate and to establish biofilm is not exclusively dependent of the type of Candida species and even strains in a blood-related infection, but it is also influenced by their interaction with host homeostasis and variations (mucosal pH shifts or nutritional changes), previous use of antibiotics, and immune system alterations (such as secondary effect of stress or immunosuppressant therapy) [35].
The I2 observed in the forest plot indicate a high heterogenic data. The I2 is a measurement of the heterogeneity that is not caused by variations in the sample size considered in each study. Therefore, this high value and also the geometry of the funnel plot indicates the possibility of major sources of variation across the studies. Some of the sources of variations can clearly be related with the differences previously described (i.e., methodology, Candida species, etc.) and consequently the pooled effect around the 80% need to be considered with caution. Several factors can be modulating this pooled effect leading to higher and/or lower values. In this context, the present meta-analysis was unable to study any correlations between clinical or epidemiological factors and mortality in patients with biofilm-related blood infections. These heterogeneity and gaps on the selected studies constitute the main drawback of our study. However, it is also well-known that the ability to establish biofilms among Candida species is an important virulence factor contributing to a more severe infection in patients [36] and it is worth to be studied. The observed heterogeneity was the leading cause to consider the effect of several variables like geographical distribution and Candida species. However, the missing information in the consulted scientific literature can be an important source of unexplained variation.
Geographical distribution of Candida biofilm-related infections
World incidence of invasive candidiasis is difficult to estimate because the criteria used for diagnosing and categorizing invasive candidiasis are quite different [6, 8, 9]. Also, most studies restricted many factors in their group set, such as the range age of patients and their health status. The present meta-analysis recollected data from diverse study sets demonstrating the Candida-related biofilm infections as a main nosocomial infection, but only 16 of 31 studies partially reported the clinical background of the patients (Table 2), such as patients suffering from immunodeficiency, receiving organ transplantations, under major surgery, or treated with cancer chemotherapy and different primary hospitalizations, and no epidemiological factors were available. Only a study realized in a tertiary care hospital of southern India reported the clinical backgrounds in adult and pediatric patients [37], evidencing central venous catheter and low weight at birth as the most prevalent risk factors in these population sets, respectively.
Generally, the number of patients in surveillance studies is very low and there are many gaps in our knowledge on the true epidemiology of invasive candidiasis in many regions of the world [19]. As expected, around 55% of our data set belonged to European studies (17/31), where the rate of biofilm-related infections varied greatly among countries showing Spain with statistical differences in the incidence of Candida-related biofilm infections in hospitalized patients in comparison with other countries. However, Cesta and colleagues recently reported Italy as the one region with a higher number of deaths caused by antibiotic-resistant bacteria and biofilm-related infections [38]. Due to European Centre for Disease Prevention and Control (ECDC) reported a spread of multi-drug resistant strains (MDR) in Italy, in particular of the bacterial species of Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii [38], it is plausible that the Candida-related biofilm incidence among hospitalized patients in Italy had been underrated. Likewise, only two and three studies in our data set belong to North and South America, respectively. All three studies of South America were indeed from Brazil, demonstrating one of the highest Candida-related biofilm incidences among hospitalized patients (91.6%). However, no further information was available in the remaining Latin-American countries with the criteria selection of the present meta-analysis.
We can notice in the meta-analysis that the values of I2, Q and other indicators also suggest a high heterogeneity within each group. It is an indicator that other factors can be involved. For example, if we consider only the articles from Italy, we can notice that the sample size in 5 of 6 studies do not considerably differ but the effect size is quite different (this will impact directly in the funnel plot geometry as presented in Fig 3). In three studies, we found a low prevalence of biofilm formation [33, 39, 40] while in other two articles we found a high prevalence of biofilm formation [41, 42]. This distribution suggests that factors quite beyond the geography are possible causes of heterogeneity within groups.
Association between different Candida species in biofilm and infections
The number of Candida species with clinical importance in humans is relatively small, more exactly, Candida albicans, Candida glabrata, Candida tropicalis, Candida parapsilosis, and Candida dubliniensis [43]. C. albicans is the most reported Candida species worldwide in different ethnic populations [34, 44–47], being responsible for the majority of oral and systemic candidiasis cases. However, there has been an increase in the number of reports about non-albicans Candida infection in the last years and even surpassing C. albicans in terms of incidence and attributable mortality [25, 31, 34, 42, 48–51]. This new scenario could be attributed to the implementation of better molecular techniques in the identification of Candida species [21, 29, 52].
Our results demonstrated C. tropicalis as the most prevalent HBF evidencing statistical dominance among Candida species. Although C. tropicalis is described as a species with normal to high biofilm-forming capacity [36], it is commonly related to infections in prosthetic joints, endodontic issues, ulcerative colitis [53–55]. C. tropicalis biofilm is characterized by chains of cells with thin, but large, amounts of extracellular matrix material with low sums of carbohydrate and protein [36, 40]. Furthermore, Silva and colleagues showed that matrix material extracted from biofilms of C. tropicalis and C. albicans contained carbohydrates, proteins, hexosamine, phosphorus and uronic acid [55]. However, hexosamine was the major component quantified in C. tropicalis biofilm (27%). C. tropicalis biofilms are described as a dense network of yeast cells with evident different filamentous morphologies [36].
After C. tropicalis, the present meta-analysis showed C. krusei and C. glabrata as the second and third most prevalent HBF among Candida species, more exactly, 52.8 and 37.6%, respectively. C. krusei is characterized by a thick multilayered biofilm of pseudohyphal forms embedded within the polymer matrix [56], being categorized with a high ability to establish biofilm [36]. Several mucosal infections and pneumonia are caused by C. krusei [23, 56]. Although C. glabrata is known to develop less biofilm, it is characterized to produce high content of both protein and carbohydrate [40, 57]. C. glabrata is commonly associated with infections among patients with total parenteral nutrition, periodontal disease, ventilator-associated and non-healing surgical wounds [58]. C. glabrata biofilms are structured on multilayers of blastospores with high cohesion among them [55]. The elucidation of these biofilm-forming abilities and properties among Candida species could provide a promising step toward the improvement of treatments.
Until this point, we have showed that Candida species and geographical distribution can be related with our data heterogeneity. The actual combination of both variables in a multiple meta-regression model as interacting variables explained more than the 50% of the global variability. The lack of clinical information and many other discussed variables are probably related, at least partially, with the remained variability. Unfortunately, as previously explained, this information is not accessible for most of the studies and constitute by itself a recommendation in further studies.
Antifungal resistance among Candida-related biofilm infections
Candida spp. infections had successfully become more difficult to treat in the last decade due to the growth of immunogenic diseases, the disproportionate use of immunosuppressive drugs, malnutrition, endocrine disorders, the widespread use of indwelling medical devices, broad-spectrum antibiotics, aging, and an increase of the number of patients among the population [36, 59]. Thus, the morbidity and mortality associated with candidiasis are still very high, even using the actual antifungal drugs [59]. The main antifungal drugs applied to Candida infections are azoles, polyenes, and echinocandins [60]. Briefly, azoles (such as fluconazole and voriconazole) block ergosterol synthesis by targeting the enzyme lanosterol 14α-demethylase and leading to an accumulation of toxic sterol pathway intermediates. While echinocandins (such as caspofungin) aim for the synthesis of 1,3-β-glucan (a cell wall component), being the ideal antifungal drug of choice in severe cases of candidemia [61, 62]. As previously referred, the rates of antifungal resistance to fluconazole, caspofungin, and voriconazole in biofilm cells surpassed planktonic cells by a factor of 4.7, 23.5, and 42.4, respectively. Despite the number of studies comparing resistance between planktonic and biofilm cells among Candida species is still scarce, these results are in agreement with the literature postulations [36, 63]. Numerous reasons are attributed to this enormous resistance against antifungal drugs in Candida-related biofilms, such as high cell density, growth rate reduction, nutrient limitation, matrix extracellular production, presence of persister (dormant and non-dividing) cells, phenotypic shift, and high sterols content on membrane cell [36, 59, 63]. So, the treatment for Candida-biofilm infections requires a comprehensive knowledge of the complex mechanisms underlying the interaction between a biofilm and its host.
Although no efficient treatment for Candida biofilms has been found yet, several promising strategies are being explored. New therapeutic targets, such as the genes involved in biofilm development and the quorum-sensing systems, are considered an alternative treatment to the currently antifungal drugs.
Conclusions
In summary, several studies on the prevalence of Candida biofilms in bloodstream infections have been published across the world, allowing some conclusions on its mortality, species, and virulence in different geographic regions. However, a lot of information is missing, such as the lack of a thorough clinical background from the patients and the diversity of the primary infections from the patients. Further studies are needed to close gaps in our understanding of the incidence of Candida biofilms and to monitor trends in antifungal resistance and species shifts.
To the authors’ best knowledge, this meta-analysis is one of the few that explored the association of biofilm production among different Candida species in bloodstream infections [64–67], using data published worldwide and adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline. Although the present meta-analysis was performed methodically, there are some limitations of this study: (1) heterogeneity exists in some subgroup and overall analyses; (2) relationship between mortality and each Candida-related biofilm species could not be assessed; and, (3) a detailed analysis of antifungal resistance in Candida biofilms was not possible. These limitations are due to a lack of sufficient published data. Therefore, early detection of biofilms and a better characterization of Candida spp. bloodstream infections should be considered, which eventually will help preserve public health resources and ultimately diminish mortality among patients.
Materials and methods
Data selection, search strategy, and study guidelines
This study was conducted following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) strategies (S1 File) [68]. Web of Science, Scopus, PubMed, and Google Scholar databases were searched for English papers using the following medical subject heading terms (MESH): “invasive candidiasis”; “bloodstream infections”; “biofilm formation”; “biofilm-related infections”; “mortality”; and, “prevalence”.
In each electronic database, a combination of MESH terms was used to conduct the search applying the following strategy (in the MEDLINE for example): ‘‘(Candida) AND (biofilm [Title/Abstract]) AND (mortality).” All studies published until 30th July of 2020 were retrieved. The articles reporting the prevalence of bloodstream infections biofilm-related, the mortality rates, and the species identification of Candida isolates were included. The references of these articles were also checked for finding additional records. The data selection was limited to human clinical isolates and studies in English. All references were compiled into a database Zotero Library and then managed using Excel.
Screening process
Duplicates were initially identified and eliminated in Zotero after entering all the recognized studies into an Excel self-created database (S2 File). All articles were assessed by two reviewers (MBA-C and FSC-M) by screening titles, abstracts, topics, and finally full texts. At each level, the reviewers independently screened the articles and finally merged their conclusions. An additional examination of the selected articles was realized by a third author (AM) focused on the homogeneity of the eligibility criteria of previous reviewers in the initial data set. Discrepancies were resolved by discussion before finalizing the records for the evaluation of eligibility criteria. In case of disagreements, the third assessor (AM) was assigned to make a final decision.
Eligibility criteria
The major inclusion criteria included reporting the rate of biofilm formation and the prevalence of biofilm-related to Candida species, including observational studies (more exactly, cohort, retrospective, and case-control studies). Furthermore, data regarding the mortality rate, the geographical location of the study set, and the use of anti-fungal agents in clinical isolates were also extracted from the studies.
All studies without information about biofilm formation or clinical Candida isolates were consequently excluded. The method to quantify biofilm biomass was not a criterion to include or exclude any paper in this meta-analysis. Concerning antifungal resistance rate, only studies that used the standard susceptibility tests according to the Clinical and Laboratory Standards Institute (CLSI) or European Committee on Antimicrobial Susceptibility Testing EUCAST were selected for the present study.
Reviews, editorials, congress or meeting abstracts, literature in languages other than English, case reports, and letters to editors were excluded from the final data set. Finally, articles without full text available, duplicate reports on different databases, and studies with unclear or missing data were also omitted.
Data extraction and quality assessment
Methodological quality assessment of the studies was performed using a checklist for necessary items as outlined in the Critical Appraisal Skills Programmed checklists [69]. For each article, a series of critical questions were asked. If the pertinent data were presented, the question was scored ‘‘yes.” If there was any doubt or no information in the study, that question was marked as ‘‘no”. A data extraction form was designed to extract the relevant characteristics of each study (S1 and S2 Files). The extracted information included the first authors’ names, time of the study, year of publication, location, sample size, biofilm formation rate, Candida species and its categorization (as C. albicans and non-albicans Candida species), the correlation between biofilm formation and antifungal resistance, and the type of biofilm. The type of biofilm was categorized as low biofilm formers (LBF), intermediate biofilm formers (IBF), and high biofilm formers (HBF). The initial two authors (MBA-C and FSC-M) extracted all data, further confirmation and final evaluation were realized by the lead authors (AM and ET).
Data analysis and statistical methods
Meta-analysis was performed using several R packages ("meta" [70], "metafor" [71], "poibin" [72], and "stringr" [73]) of R version 3.4.3 [74] and RStudio version 1.3.1073 [75] (S3 File). The rate of biofilm formation was computed, and values were reported with confidence intervals (CI) of 95%. The heterogeneity was assessed by the Cochrane Q and I2 tests. The I2 metric indicates the amount of heterogeneity that is not related with sampling size variation. Moreover, it is also independent of the number of studies included in the meta-analysis (in contrast to the Cochrane Q metric). Considering the heterogeneity indices, the random-effects model was then used for meta-analysis of the selected studies, and the Freeman-Tukey transformation was also applied to calculate the pooled frequencies. To estimate the between-study variance in a random-effects model we use tau-squared, and its square root is the estimated standard deviation of underlying effects across studies. Subgroup analyses were conducted based on the type of biofilm, biofilm-related species, geographical regions, and antifungal resistance rates. Outliers’ analysis was done with the Baujat diagram, while quantitative Egger weighted regression test and Funnel plot were used to evaluate the eventual existence of publication bias. In statistical analysis, p-values <0.05 were considered as significant statistical results. We used the multiple meta-regression analysis with the "metareg" function from "meta" to explore the contribution to model heterogeneity of several variables. In this approach, the maximum-likelihood method was used.
Supporting information
S1 Fig. Forest plot of the meta-analysis of the prevalence of high biofilm producers in Candida spp. isolates.
https://doi.org/10.1371/journal.pone.0263522.s001
(TIF)
S2 Fig. Forest plot of the meta-analysis of the prevalence of intermediate biofilm producers in Candida spp. isolates.
https://doi.org/10.1371/journal.pone.0263522.s002
(TIF)
S3 Fig. Forest plot of the meta-analysis of the prevalence of low biofilm producers in Candida spp. isolates.
https://doi.org/10.1371/journal.pone.0263522.s003
(TIF)
S1 Table. Subgroup analysis between different Candida species and biofilm-forming capability.
https://doi.org/10.1371/journal.pone.0263522.s004
(DOCX)
S1 File. The PRISMA statement for reporting meta-analysis of the present study.
https://doi.org/10.1371/journal.pone.0263522.s005
(DOCX)
S2 File. The databases of the present study for metanalyses process.
https://doi.org/10.1371/journal.pone.0263522.s006
(XLSX)
S3 File. The R-code used in the present meta-analysis.
https://doi.org/10.1371/journal.pone.0263522.s007
(TXT)
Acknowledgments
A special recognition deserves all colleagues of the Microbiology Institute of USFQ and COCIBA, as well as the Research Office of Universidad San Francisco de Quito.
References
- 1. Tsay S V, Mu Y, Williams S, Epson E, Nadle J, Bamberg WM, et al. Burden of Candidemia in the United States, 2017. Clin Infect Dis. 2020;71: e449–e453. pmid:32107534
- 2. Thompson A, Davies LC, Liao C-T, da Fonseca DM, Griffiths JS, Andrews R, et al. The protective effect of inflammatory monocytes during systemic C. albicans infection is dependent on collaboration between C-type lectin-like receptors. Hohl TM, editor. PLOS Pathog. 2019;15: e1007850. pmid:31242262
- 3. Bassetti M, Giacobbe DR, Vena A, Trucchi C, Ansaldi F, Antonelli M, et al. Incidence and outcome of invasive candidiasis in intensive care units (icus) in europe: Results of the eucandicu project. Crit Care. 2019;23: 1–7. pmid:30606235
- 4. Hsueh PR, Graybill JR, Playford EG, Watcharananan SP, Oh MD, Ja’alam K, et al. Consensus statement on the management of invasive candidiasis in Intensive Care Units in the Asia-Pacific Region. Int J Antimicrob Agents. 2009;34: 205–209. pmid:19409759
- 5. Nucci M, Queiroz-Telles F, Alvarado-Matute T, Tiraboschi IN, Cortes J, Zurita J, et al. Epidemiology of Candidemia in Latin America: A Laboratory-Based Survey. PLoS One. 2013;8: e59373. pmid:23527176
- 6. Colombo AL, Cortes JA, Zurita J, Guzman-Blanco M, Alvarado Matute T, de Queiroz Telles F, et al. Recomendaciones para el diagnóstico de la candidemia en América Latina. Rev Iberoam Micol. 2013;30: 150–157. pmid:23764553
- 7. Clancy CJ, Nguyen MH. Finding the missing 50% of invasive candidiasis: How nonculture diagnostics will improve understanding of disease spectrum and transform patient care. Clin Infect Dis. 2013;56: 1284–1292. pmid:23315320
- 8. Lagunes L, Rello J. Invasive candidiasis: From mycobiome to infection, therapy, and prevention. European Journal of Clinical Microbiology and Infectious Diseases. Springer Verlag; 2016. pp. 1221–1226. pmid:27146877
- 9. Pappas PG, Kauffman CA, Andes DR, Clancy CJ, Marr KA, Ostrosky-Zeichner L, et al. Clinical Practice Guideline for the Management of Candidiasis: 2016 Update by the Infectious Diseases Society of America. Clin Infect Dis. 2015;62: e1–e50. pmid:26679628
- 10. Guembe M, Guinea J, Marcos-Zambrano L, Fernández-Cruz A, Peláez T, Muñoz P, et al. Is Biofilm Production a Predictor of Catheter-Related Candidemia? Med Mycol. 2014;52: 407–410. pmid:24782103
- 11. Chandra J, Mukherjee PK. Candida Biofilms: Development, Architecture, and Resistance. Microbiol Spectr. 2015;3: 1–14. pmid:26350306
- 12. Nobile CJ, Johnson AD. Candida albicans Biofilms and Human Disease. Annu Rev Microbiol. 2015;69: 71–92. pmid:26488273
- 13. Polke M, Hube B, Jacobsen ID. Candida survival strategies. Advances in Applied Microbiology. Elsevier Ltd; 2015. pmid:25911234
- 14. Johnson CJ, Cabezas-Olcoz J, Kernien JF, Wang SX, Beebe DJ, Huttenlocher A, et al. The Extracellular Matrix of Candida albicans Biofilms Impairs Formation of Neutrophil Extracellular Traps. PLoS Pathog. 2016;12: 1–23. pmid:27622514
- 15. Branchini ML, Pfaller MA, Rhine-Chalberg J, Frempong T, Isenberg HD. Genotypic Variation and Slime Production among Blood and Catheter Isolates of Candida parapsilosis. J Clin Microbiol. 1994;32: 452–456. pmid:8150956
- 16. Monfredini PM, Souza ACR, Cavalheiro RP, Siqueira RA, Colombo AL. Clinical impact of Candida spp. biofilm production in a cohort of patients with candidemia. Med Mycol. 2018;56: 803–808. pmid:29228246
- 17. Vitális E, Nagy F, Tóth Z, Forgács L, Bozó A, Kardos G, et al. Candida biofilm production is associated with higher mortality in patients with candidaemia. Mycoses. 2020;63: 352–360. pmid:31943428
- 18. Xiao G, Liao W, Zhang Y, Luo X, Zhang C, Li G, et al. Analysis of fungal bloodstream infection in intensive care units in the Meizhou region of China: Species distribution and resistance and the risk factors for patient mortality. BMC Infect Dis. 2020;20: 599. pmid:32795259
- 19. Quindós G, Marcos-Arias C, San-Millán R, Mateo E, Eraso E. The continuous changes in the aetiology and epidemiology of invasive candidiasis: from familiar Candida albicans to multiresistant Candida auris. International Microbiology. Springer; 2018. pp. 107–119. pmid:30810955
- 20. Sahal G, Bilkay IS. Distribution of clinical isolates of Candida spp. and antifungal susceptibility of high biofilm-forming Candida isolates. Rev Soc Bras Med Trop. 2018;51: 644–650. pmid:30304271
- 21. Devi AR, Hymavathi R, Mounika G. Candida Species Isolation, Identification and Biofilm Detection at a Tertiary Care Hospital. Int J Contemp Med Res. 2019;6: D6–D9.
- 22. Marak MB, Dhanashree B. Antifungal Susceptibility and Biofilm Production of Candida spp. Isolated from Clinical Samples. 2018. pmid:30405717
- 23. Mohandas V, Ballal M. Distribution of Candida Species in different clinical samples and their virulence: Biofilm formation, proteinase and phospholipase production: A study on hospitalized patients in Southern India. J Glob Infect Dis. 2011;3: 4–8. pmid:21572601
- 24. Villar-Vidal M, Marcos-Arias C, Eraso E, Quindós G. Variation in biofilm formation among blood and oral isolates of Candida albicans and Candida dubliniensis. Enferm Infecc Microbiol Clin. 2011;29: 660–665. pmid:21899928
- 25. Tulasidas S, Rao P, Bhat S, Manipura R. Infection and Drug Resistance Dovepress A study on biofilm production and antifungal drug resistance among candida species from vulvovaginal and bloodstream infections. 2018. pmid:30538510
- 26. Owen MK, Clenney TL. Management of vaginitis. Am Fam Physician. 2004;70. pmid:15606061
- 27. Høiby N, Bjarnsholt T, Moser C, Bassi GL, Coenye T, Donelli G, et al. ESCMID* guideline for the diagnosis and treatment of biofilm infections 2014. Clin Microbiol Infect. 2015;21: S1–S25. pmid:25596784
- 28. Bouza E, Alcalá L, Muñoz P, Martín-Rabadán P, Guembe M, Rodríguez-Créixems M. Can microbiologists help to assess catheter involvement in candidaemic patients before removal? Clin Microbiol Infect. 2013;19. pmid:23231412
- 29. Carvalho A, Costa-De-Oliveira S, Martins ML, Pina-Vaz C, Rodrigues AG, Ludovico P, et al. Multiplex PCR identification of eight clinically relevant Candida species. Med Mycol. 2007;45: 619–27. pmid:17885953
- 30. Hirano R, Sakamoto Y, Kudo K, Ohnishi M. Retrospective analysis of mortality and Candida isolates of 75 patients with candidemia: A single hospital experience. Infect Drug Resist. 2015;8: 199–205. pmid:26185460
- 31. Alkharashi N, Aljohani S, Layqah L, Masuadi E, Baharoon W, Al-Jahdali H, et al. Candida Bloodstream Infection: Changing Pattern of Occurrence and Antifungal Susceptibility over 10 Years in a Tertiary Care Saudi Hospital. Can J Infect Dis Med Microbiol. 2019;2019. pmid:31929847
- 32. Ghrenassia E, Mokart D, Mayaux J, Demoule A, Rezine I, Kerhuel L, et al. Candidemia in critically ill immunocompromised patients: report of a retrospective multicenter cohort study. Ann Intensive Care. 2019;9: 62. pmid:31161475
- 33. Tumbarello M, Posteraro B, Trecarichi EM, Fiori B, Rossi M, Porta R, et al. Biofilm production by Candida species and inadequate antifungal therapy as predictors of mortality for patients with candidemia. J Clin Microbiol. 2007;45: 1843–1850. pmid:17460052
- 34. Rajendran R, Sherry L, Nile CJ, Sherriff A, Johnson EM, Hanson MF, et al. Biofilm formation is a risk factor for mortality in patients with Candida albicans bloodstream infection-Scotland, 2012–2013. Clin Microbiol Infect. 2016;22: 87–93. pmid:26432192
- 35. Lohse MB, Gulati M, Johnson AD, Nobile CJ. Development and regulation of single-and multi-species Candida albicans biofilms. Nat Rev Microbiol. 2018;16: 19–31. pmid:29062072
- 36. Silva S, Rodrigues CF, Araújo D, Rodrigues ME, Henriques M. Candida species biofilms’ antifungal resistance. J Fungi. 2017;3: 8. pmid:29371527
- 37. Banerjee B, Saldanha Dominic RM, Baliga S. Clinico-microbiological study of candidemia in a tertiary care hospital of southern part of India. Iran J Microbiol. 2015;7: 55. pmid:26644875
- 38. Cesta N, Di Luca M, Corbellino M, Tavio M, Galli M, Andreoni M. Bacteriophage therapy: An overview and the position of Italian society of infectious and tropical diseases. Infez Med. 2020;28: 322–331. pmid:32920567
- 39. Tortorano AM, Prigitano A, Lazzarini C, Passera M, Deiana ML, Cavinato S, et al. A 1-year prospective survey of candidemia in Italy and changing epidemiology over one decade. Infection. 2013;41: 655–662. pmid:23559357
- 40. Tumbarello M, Fiori B, Trecarichi EM, Posteraro P, Losito AR, de Luca A, et al. Risk factors and outcomes of candidemia caused by biofilm-forming isolates in a tertiary care hospital. PLoS One. 2012;7: 1–9. pmid:22479431
- 41. Prigitano A, Dho G, Lazzarini C, Ossi C, Cavanna C, Tortorano AM. Biofilm production by Candida isolates from a survey of invasive fungal infections in italian intensive care units. J Chemother. 2012;24: 61–63. pmid:22546727
- 42. Soldini S, Posteraro B, Vella A, De Carolis E, Borghi E, Falleni M, et al. Microbiologic and clinical characteristics of biofilm-forming Candida parapsilosis isolates associated with fungaemia and their impact on mortality *. 2018. pmid:29133157
- 43. McManus BA, Coleman DC. Molecular epidemiology, phylogeny and evolution of Candida albicans. Infection, Genetics and Evolution. Elsevier; 2014. pp. 166–178. pmid:24269341
- 44. Toda M, Williams SR, Berkow EL, Farley MM, Harrison LH, Bonner L, et al. Population-Based Active Surveillance for Culture-Confirmed Candidemia—Four Sites, United States, 2012–2016. MMWR Surveill Summ. 2019;68: 1–15. pmid:31557145
- 45. Brunetti G, Navazio AS, Giuliani A, Giordano A, Proli EM, Antonelli G, et al. Candida blood stream infections observed between 2011 and 2016 in a large Italian University Hospital: A time-based retrospective analysis on epidemiology, biofilm production, antifungal agents consumption and drug-susceptibility. Cortegiani A, editor. PLoS One. 2019;14: e0224678. pmid:31697722
- 46. Yen Tan T, Ling Tan A. A Retrospective Analysis of Antifungal Susceptibilities of Candida Bloodstream Isolates From Singapore Hospitals. 2008.
- 47. Barber K, Wagner J, Miller J, Lewis E, Stover K. Impact of Obesity in Patients with Candida Bloodstream Infections: A Retrospective Cohort Study. Infect Dis Ther. 2020;9. pmid:32062851
- 48. Raja NS. Epidemiology, risk factors, treatment and outcome of Candida bloodstream infections because of Candida albicans and Candida non‐albicans in two district general hospitals in the United Kingdom. Int J Clin Pract. 2021;75. pmid:32869497
- 49. Al-Musawi TS, Alkhalifa WA, Alasaker NA, Rahman JU, Alnimr AM. A seven-year surveillance of Candida bloodstream infection at a university hospital in KSA. J Taibah Univ Med Sci. 2020. pmid:33897322
- 50. Bhatt M, Sarangi G, Paty BP, Mohapatra D, Chayani N, Mahapatra A, et al. Biofilm as a virulence marker in Candida species in Nosocomial blood stream infection and its correlation with antifungal resistance. Indian J Med Microbiol. 2015;33: S112–S114. pmid:25657126
- 51. Zhang W, Song X, Wu H, Zheng R. Epidemiology, species distribution, and predictive factors for mortality of candidemia in adult surgical patients. BMC Infect Dis. 2020;20: 506. pmid:32660641
- 52. Romeo O, Scordino F, Pernice I, Lo Passo C, Criseo G. A multiplex PCR protocol for rapid identification of Candida glabrata and its phylogenetically related species Candida nivariensis and Candida bracarensis. J Microbiol Methods. 2009;79: 117–120. pmid:19635503
- 53. Mansur AJ, Safi J Jr, Markus M, Aiello VD, Grinberg M, Pomerantzeff PM. Late failure of surgical treatment for bioprosthetic valve endocarditis due to Candida tropicalis. Clin Infect Dis. 1996;22: 380–381. pmid:8838207
- 54. Negri M, Silva S, Henriques M, Oliveira R. Insights into Candida tropicalis nosocomial infections and virulence factors. Eur J Clin Microbiol Infect Dis. 2012;31: 1399–1412. pmid:22037823
- 55. Silva S, Negri M, Henriques M, Oliveira R, Williams DW, Azeredo J. Candida glabrata, Candida parapsilosis and Candida tropicalis: Biology, epidemiology, pathogenicity and antifungal resistance. FEMS Microbiol Rev. 2012;36: 288–305. pmid:21569057
- 56. Pannanusorn S, Fernandez V, Römling U. Prevalence of biofilm formation in clinical isolates of Candida species causing bloodstream infection. Mycoses. 2013;56: 264–272. pmid:23113805
- 57. Hawser SP, Douglas LJ. Resistance of Candida albicans biofilms to antifungal agents in vitro. Antimicrob Agents Chemother. 1995;39: 2128–2131. pmid:8540729
- 58. Rodrigues CF, Silva S, Henriques M. Candida glabrata: A review of its features and resistance. Eur J Clin Microbiol Infect Dis. 2014;33: 673–688. pmid:24249283
- 59. Alves R, Barata-Antunes C, Casal M, Brown AJP, van Dijck P, Paiva S. Adapting to survive: How Candida overcomes host-imposed constraints during human colonization. PLoS Pathog. 2020;16: e1008478. pmid:32437438
- 60. Santos GC d. O, Vasconcelos CC, Lopes AJO, Cartágenes M do S d. S, Filho AKDB, do Nascimento FRF, et al. Candida infections and therapeutic strategies: Mechanisms of action for traditional and alternative agents. Front Microbiol. 2018;9: 1–23. pmid:29403456
- 61. White TC, Marr KA, Bowden RA. Clinical, cellular, and molecular factors that contribute to antifungal drug resistance. Clin Microbiol Rev. 1998;11: 382–402. pmid:9564569
- 62. Pappas PG, Kauffman CA, Andes DR, Clancy CJ, Marr KA, Ostrosky-Zeichner L, 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–e50. pmid:26679628
- 63. Cavalheiro M, Teixeira MC. Candida Biofilms: Threats, challenges, and promising strategies. Front Med. 2018;5: 1–15. pmid:29487851
- 64. Pammi M, Holland L, Butler G, Gacser A, Bliss JM. Candida parapsilosis Is a Significant Neonatal Pathogen: A Systematic Review and Meta-Analysis. Pediatr Infect Dis J. 2013;32: e206–e216. pmid:23340551
- 65. Buehler SS, Madison B, Snyder SR, Derzon JH, Cornish NE, Saubolle MA, et al. Effectiveness of practices to increase timeliness of providing targeted therapy for inpatients with bloodstream infections: A laboratory medicine best practices systematic review and meta-analysis. Clin Microbiol Rev. 2015;29: 59–103. pmid:26598385
- 66. Kobayashi T, Marra AR, Schweizer ML, Eyck P Ten, Wu C, Alzunitan M, et al. Impact of infectious disease consultation in patients with candidemia: A retrospective study, systematic literature review, and meta-analysis. Open Forum Infect Dis. 2020;7: 1–11. pmid:32904995
- 67. Pinto H, Simões M, Borges A. Prevalence and impact of biofilms on bloodstream and urinary tract infections: A systematic review and meta-analysis. Antibiotics. 2021;10: 1–24. pmid:34356749
- 68. Selcuk AA. A Guide for Systematic Reviews: PRISMA. Turkish Arch Otorhinolaryngol. 2019;57: 57–58. pmid:31049257
- 69. Zeng X, Zhang Y, Kwong JSW, Zhang C, Li S, Sun F, et al. The methodological quality assessment tools for preclinical and clinical studies, systematic review and meta-analysis, and clinical practice guideline: A systematic review. Journal of Evidence-Based Medicine. Blackwell Publishing; 2015. pp. 2–10. pmid:25594108
- 70. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with R: A practical tutorial. Evid Based Ment Health. 2019;22: 153–160. pmid:31563865
- 71. Viechtbauer W. Conducting Meta-Analyses in R with the metafor Package. J Stat Softw. 2010;36: 1–48.
- 72. Hong Y. On computing the distribution function for the Poisson binomial distribution. Comput Stat Data Anal. 2013;59: 41–51. Available: https://econpapers.repec.org/RePEc:eee:csdana:v:59:y:2013:i:c:p:41-51
- 73.
Wickham H. stringr: Simple, Consistent Wrappers for Common String Operations. 2021. p. http://stringr.tidyverse.org, https://github.com/t.
- 74.
R Core Team. R: A language and environment for statistical computing. Vienna (Austria): R Foundation for Statistical Computing; 2021. Available: https://www.r-project.org/.
- 75.
RStudio Team. RStudio: Integrated Development for R. Boston (USA): RStudio, Inc.; 2021. Available: http://www.rstudio.com/