Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
SUPPLEMENT ARTICLE Characteristics, Clinical Relevance, and the Role of Echinocandins in Fungal–Bacterial Interactions Marios Arvanitis1,2,3 and Eleftherios Mylonakis1,2 1 Infectious Diseases Division, Rhode Island Hospital, and 2Warren Alpert Medical School of Brown University, Providence, Rhode Island; and 3Internal Medicine Department, Boston Medical Center, Massachusetts Fungal–bacterial interactions are common in the environment. The interactions between invasive fungi (eg, Candida species and Aspergillus species) and pathogenic bacteria can be particularly significant in the outcome of human infections. Study of these interactions in vivo using murine or invertebrate models, such as Caenorhabditis elegans or Galleria mellonella, has been very helpful in increasing our understanding of the pathogenesis of mixed infections and in identifying ways to use this between-kingdom interplay to our advantage. Based on their effect against fungal biofilms and their immunomodulatory properties, the newer class of antifungal agents, known as echinocandins, has the potential to be useful in polymicrobial infections and in high-risk complex infections such as ventilator-associated pneumonia or sepsis where colonization by fungi can lead to worse outcomes. Keywords. candidiasis; aspergillosis; micafungin; caspofungin; Caenorhabditis elegans. The interactions between different species have been an essential feature of life and, in their struggle for survival, the various living organisms interact with each other in a variety of ways. Often, this interaction leads to synergistic cooperation, from which both species will eventually benefit, whereas in other circumstances the interaction becomes antagonistic, in which case the less powerful part must either adapt, by developing features that will help improve its chances of survival, or die. This second type of interaction has had profound implications in life as we know it today, as it has been one of the driving forces of natural selection [1]. These interspecies interactions become particularly important in the microscopic level among monocellular organisms such as bacteria and fungi. Correspondence: Eleftherios Mylonakis, MD, PhD, FIDSA, Infectious Diseases Department, Rhode Island Hospital, 593 Eddy St, 3rd Flr, Ste 328/330, Providence, RI 02903 ([email protected]). Clinical Infectious Diseases® 2015;61(S6):S630–4 © The Author 2015. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: [email protected]. DOI: 10.1093/cid/civ816 S630 • CID 2015:61 (Suppl 6) • Arvanitis and Mylonakis FUNGAL–BACTERIAL INTERACTIONS Interactions between prokaryotes and eukaryotes are abundant in nature and have been described in detail by several in vitro experiments. In some circumstances, these interactions can be beneficial for both life forms. For example, researchers recently found that Candida albicans biofilms can support and promote growth of certain anaerobic bacteria by protecting them from ambient toxic conditions. To enhance this phenomenon, the anaerobes promote aggregation of yeast cells into mini-biofilms [2]. A similar relationship between C. albicans and Streptococcus mutans was identified in another article. The presence of C. albicans increases the production of exopolysaccharides and thus promotes S. mutans biofilm formation, while also inducing the expression of several virulence factors of the bacteria [3]. Similarly, ethanol production by C. albicans has been shown to promote Pseudomonas aeruginosa biofilm development through stimulation of the diguanylatecyclase WspR [4]. However, not all interactions between bacteria and fungi are synergistic. Indeed, physical interactions between Candida species and prokaryotes in which the bacteria develop biofilms on the yeast hyphae, destroying them in the process, are well known [5]. In addition, several secretory molecules produced by fungi have been proven to be bactericidal, with, of course, the most widely known example being the discovery of the antimicrobial activities of penicillin produced by the fungus Penicillium rubens [6]. The research for identification of other fungal metabolites with antibacterial effects continues to date [7]. Similarly, bacteria are able to produce secretory molecules that inhibit the growth of fungi. For example, volatile compounds produced by P. aeruginosa [8] or Burkholderia tropica [9] have been shown to inhibit cytopathogenic fungi of the genus Fusarium. Finally, changes in environmental conditions caused by certain bacterial species can impact the growth of fungi. For example, lactobacilli are known to cause a decrease in the pH of human surfaces, thus leading to elimination of pathogenic fungi such as Candida species [10]. While these between-kingdom effects have clearly been elucidated by in vitro experiments and environmental observations, the relevance and extent of these relationships to human infections have only recently started to be realized. Fungal species are abundant in nature and thus are frequent colonizers of the gastrointestinal tract and other parts of multicellular organisms. Therefore, bacterial pathogens that invade these organs are often faced with the presence of these fungi. Furthermore, mixed bacterial and fungal infections can occur in high-risk patients [11]. Under those circumstances, the outcome of the infection frequently depends on the character of fungal–bacterial interactions. Experiments on invertebrate model hosts have been important in studying the pathogenesis of those mixed polymicrobial infections. Invertebrate models are small multicellular organisms that are easy to breed and use in a microbiology laboratory and can provide useful information regarding human infections, including microbial virulence and host immune responses, thanks to the similarities of their immune system to the innate immune response used by humans against infectious pathogens. The worm Caenorhabditis elegans [12, 13] and the insect Galleria mellonella [14] are among the most widely used invertebrate models in experimental models of infection pathogenesis and microbial virulence. In a pivotal article, Peleg et al used C. elegans to study the in vivo interactions between C. albicans and Acinetobacter baumannii [15]. The investigators discovered that A. baumannii is able to inhibit yeast filaments, thus decreasing its virulence and partially protecting worms from a lethal C. albicans gut infection. On the other hand, when able to develop a quorum, the fungus is able to respond to this offensive move by inhibiting A. baumannii growth through secretion of the quorum-sensing molecule farnesol. In a subsequent article, the investigators advanced this experiment by assessing the interactions between C. albicans and Salmonella enterica. They found that similar to the previous experiment, Salmonella species are able to inhibit C. albicans filamentation in the gut of C. elegans via a secretory molecule produced by the bacterium. Furthermore, the decrease in C. albicans virulence, while most potent against filaments, also extends to yeast cells as well as C. albicans biofilms [16]. Kim et al were able to identify this secreted molecule as SopB, which is an effector of a type II secretion system produced by Salmonella species [17]. A similar interaction has been found between gram-positive bacteria and C. albicans. Specifically, researchers using a C. elegans gut infection model showed that coinfection with Enterococcus faecalis and C. albicans was associated with reduced virulence compared with infection with either species alone [18]. Conversely, there are some data suggesting that coinfection of C. elegans with Saccharomyces species and Acinetobacter species may result in increased bacterial virulence through the production of ethanol by the fungus [19]. An often underappreciated type of polymicrobial interaction within a multicellular host relies on the effect of different pathogens on host immune defenses. These effects have been clearly demonstrated in murine experiments. For example, using a murine model of lung infection, investigators showed that colonization of the respiratory tract by C. albicans protected mice against P. aeruginosa–induced lung damage through recruitment of natural killer cells, dendritic cells, and innate lymphoid cells and the secretion of the cytokine interleukin 22 [20]. A different relationship between C. albicans and pneumonia was found in a study on rats, which showed that Candida species colonization induced a Th1–Th17 immune response with high circulating interferon-γ levels and favored the development of bacterial pneumonia, whereas antifungal treatment was able to reverse this phenomenon and decrease the incidence of bacterial infection [21]. In parallel, coinfecting mice with C. albicans and nonfermenting gram-negative commensal bacteria results in exacerbation of the fungal infection through interferon-γ secretion [22]. Similarly, in a murine model, polymicrobial peritonitis caused by C. albicans and Staphylococcus aureus results in increased virulence and mortality compared with singlepathogen infection. This increased virulence is mediated by innate proinflammatory cytokines, leading to increased inflammation and end-organ damage [23]. EFFECT OF EUKARYOTE–PROKARYOTE INTERACTIONS ON HUMAN INFECTIONS Taken together, these observations imply that the interactions between bacteria and fungi in a living host are diverse, and their outcome may often depend on the specific environmental conditions, the potency and characteristics of the host immune response, and the species of the microorganisms involved (Table 1). Similar to what has been described from in vitro studies and in vivo models of infection, fungal–bacterial interactions Echinocandins and Fungal-Bacterial Interactions • CID 2015:61 (Suppl 6) • S631 Table 1. Types of Fungal–Bacterial Interactions Example of Antagonistic Interactions Type of Interaction Example of Synergistic Interactions Physical interaction Formation of fungal biofilms that encase and protect bacterial cells [2] Formation of bacterial biofilms on fungal hyphae leading to their destruction [5] Secretory molecules Formation of ethanol by Saccharomyces spp leads to increased virulence of Acinetobacter baumannii [4] None Pseudomonas spp phenazines inhibit Candida spp hyphae [7] Gram-negative nonfermenting bacteria exacerbate infection of the gut of mice by Candida albicans through interferon-γ [22] Candida albicans recruits lymphoid cells in the respiratory tract of mice, protecting them from subsequent Pseudomonas aeruginosa pneumonia [20] Changes in the environment Alteration of host immune response Lactobacillus spp can lower the environmental pH, thus eliminating Candida spp [10] may play an important role in the outcome of certain human infections. This becomes particularly significant in the case of infections in organs where fungi can become colonizers and thus interact with invading bacterial pathogens. One such type of infection is ventilator-associated pneumonia (VAP). In populations of critically ill, intubated patients, Candida species frequently become colonizers of the respiratory tract via transmission through the endotracheal tube, with large trials estimating the prevalence of Candida species colonization up to 16% [24]. Moreover, the respirator is used by both bacteria and fungi as an abiotic surface to promote biofilm formation [25]. In this setting, it is anticipated that the interplay between the yeast and bacterial pathogens could be an essential determinant of the outcome of pneumonia. Indeed, several observational studies have addressed this hypothesis. In the first study of its kind, Azoulay et al found that in hospitalized critically ill patients, colonization of the respiratory tract with Candida species was an independent risk factor for pneumonia, with a greatest risk for Pseudomonas species pneumonia [26]. In a different study, researchers investigated the importance of the presence of Candida species in cultures of the respiratory tract of patients with suspected VAP and found increased intensive care unit (ICU) stay and 28-day mortality in patients who were Candida colonized [27]. These findings were supported by 2 retrospective studies which showed that Candida species colonization is associated with prolonged hospital stay and higher in-hospital mortality in VAP [28, 29]. On the other hand, a more recent trial assessed 200 patients with VAP from a tertiary referral center S632 • CID 2015:61 (Suppl 6) • Arvanitis and Mylonakis within a 5-year period and found that colonized patients had increased ICU stay but not higher 30-day mortality compared with controls [30]. Taken in their totality, these studies suggest that Candida species colonization may have a detrimental effect on the outcome of patients with VAP. Notably, studies on the implications of mixed bacterial– fungal systemic infections are generally lacking. One older study that evaluated the clinical course of mixed bacterial–fungal infections found that the outcome of bacteremia is worse when Candida species are present in the blood concurrently with the bacteria compared to single-bacteria infection [31]. This study, however, as with all the other human studies of mixed infections, is limited by its observational nature, thus being unable to establish causation. Indeed, the worse outcomes observed in patients who are coinfected with fungi or are colonized by Candida species may be attributed to other comorbidities. THE ROLE OF ANTIFUNGAL AGENTS AND FUTURE DIRECTIONS Although randomized studies to directly address the comparative outcomes of mixed bacterial–fungal vs single-pathogen infections have profound ethical limitations, indirect trials that could evaluate the effect of antifungal treatment in colonized individuals could be performed without similar implications. As suggested by the studies described above, in most cases of human bacterial infections, fungal presence is associated with increased morbidity. Therefore, the intuitive thought is that the use of antifungal agents would be beneficial by preventing bacterial infections or even improving outcomes in actively infected individuals who are colonized or have a mixed infection. Surprisingly, this has not been extensively studied. In a doubleblind randomized trial, Jacobs et al evaluated 71 patients with septic shock who were randomized to receive daily fluconazole therapy vs placebo [32]. Interestingly, the patients who received fluconazole had significantly improved 30-day survival (78% vs 46%), with the benefit being more pronounced in patients with intra-abdominal infections. Nevertheless, due to the study design, it was unclear whether this could be attributed to direct antifungal effect of fluconazole or to immunomodulatory properties of the agent. Conversely, a more recent pilot randomized trial failed to support the use of antifungal agents to treat Candida species colonization in patients with a clinical suspicion for VAP [33]. However, due to the relatively small number of participants (30 patients per group), the study may have been underpowered to show statistical differences in clinical outcomes. Therefore, further assessment of the role of antifungal agents in mixed infections is warranted. Of all the antifungal agents available to date, the relatively new class of echinocandins holds a promising potential against polymicrobial infections. Contrary to the azole and polyene classes of antifungal agents, echinocandins seem to be active against fungal biofilms. Indeed, several in vitro studies [34, 35], as well as reports from mammalian infection models [36], show that different echinocandins such as caspofungin, anidulafungin, and micafungin can eliminate mature Candida species biofilms [37]. Therefore, this class of antifungal agents may be a particularly attractive option in mixed infections, especially those that develop on abiotic surfaces (ie, VAP or catheter-related infections). An additional advantage of the echinocandins that could prove to be useful in polymicrobial infections is their immunomodulatory ability. Several studies have shown that antifungal agents such as caspofungin and micafungin are able to alter the host immune response. Specifically, using the invertebrate model host G. mellonella, Kelly et al showed that preexposure to caspofungin can prime the immune response of the larvae and increase their survival after a subsequent lethal C. albicans inoculation. Further, this immune potentiation was nonspecific, as it also resulted in prolonged survival of larvae that were subsequently infected with S. aureus [38]. Moreover, caspofungin and micafungin have been shown to prime the human neutrophil response to Aspergillus species hyphae by inducing exposure of β-glucan on the hyphal surface [39]. Finally, different investigators showed that caspofungin can enhance the activity of impaired neutrophils from renal transplant recipients against Candida species [40]. These effects on the host immune response have the potential to be significant in mixed infections, both by a nonspecific potentiation of innate immune effectors that can act against both bacteria and fungi and by increased exposure of mold and yeast hyphae to lymphoid cells, which could be impaired in some polymicrobial infections [41]. Based on these findings, one could argue that echinocandins have the potential to be of use in clinical practice to not only treat active invasive fungal infections, but also to improve survival in mixed infections or in critically ill septic individuals who are colonized by fungal species. On the basis of this assumption, a double-blind randomized trial recently recruited 222 ventilated ICU patients who were receiving antibiotics and randomized them to receive prophylactic or preemptive caspofungin vs placebo [42]. The results showed that caspofungin tended to reduce the incidence of candidemia in these patients, but the results did not reach statistical significance. To give a clearer answer to the question of whether echinocandins can help in polymicrobial infections, a large multicenter randomized trial that compares micafungin and placebo in Candida-colonized patients with sepsis in the ICU setting is currently under way [43]. CONCLUSIONS Interactions between fungal and bacterial species are abundant in nature and can be relevant in human infections. Study of the pathogenesis of these interactions may provide an insight on how to treat complicated mixed species infections in critically ill individuals. With their immunomodulatory and antibiofilm properties, the newer class of echinocandins has a theoretical advantage over older antifungal agents in polymicrobial infections and in fungal-colonized patients with sepsis. Whether this holds true in clinical practice remains to be seen in future randomized trials. Notes Supplement sponsorship. This article appears as part of the supplement “Advances and New Directions for Echinocandins,” sponsored by Astellas Pharma Global Development, Inc. Potential conflict of interest. Both authors: No potential conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. Plata G, Henry CS, Vitkup D. Long-term phenotypic evolution of bacteria. Nature 2015; 517:369–72. 2. Fox EP, Cowley ES, Nobile CJ, Hartooni N, Newman DK, Johnson AD. Anaerobic bacteria grow within Candida albicans biofilms and induce biofilm formation in suspension cultures. Curr Biol 2014; 24:2411–6. 3. Falsetta ML, Klein MI, Colonne PM, et al. Symbiotic relationship between Streptococcus mutans and Candida albicans synergizes virulence of plaque biofilms in vivo. Infect Immun 2014; 82:1968–81. 4. Chen AI, Dolben EF, Okegbe C, et al. Candida albicans ethanol stimulates Pseudomonas aeruginosa WspR-controlled biofilm formation as part of a cyclic relationship involving phenazines. PLoS Pathog 2014; 10:e1004480. 5. Hogan DA, Kolter R. Pseudomonas–Candida interactions: an ecological role for virulence factors. Science 2002; 296:2229–32. 6. Houbraken J, Frisvad JC, Samson RA. Fleming’s penicillin producing strain is not Penicillium chrysogenum but P. rubens. IMA Fungus 2011; 2:87–95. 7. Nishanth Kumar S, Nisha GV, Sudaresan A, et al. Synergistic activity of phenazines isolated from Pseudomonas aeruginosa in combination with azoles against Candida species. Med Mycol 2014; 52:482–90. 8. Cordero P, Principe A, Jofre E, Mori G, Fischer S. Inhibition of the phytopathogenic fungus Fusarium proliferatum by volatile compounds produced by Pseudomonas. Arch Microbiol 2014; 196:803–9. 9. Elshafie HS, Camele I, Racioppi R, Scrano L, Iacobellis NS, Bufo SA. In vitro antifungal activity of Burkholderia gladioli pv. agaricicola against some phytopathogenic fungi. Int J Mol Sci 2012; 13:16291–302. 10. Boris S, Barbes C. Role played by lactobacilli in controlling the population of vaginal pathogens. Microbes Infect 2000; 2:543–6. 11. Kaufman D, Fairchild KD. Clinical microbiology of bacterial and fungal sepsis in very-low-birth-weight infants. Clin Microbiol Rev 2004; 17:638–80, table of contents. 12. Schulenburg H, Kurz CL, Ewbank JJ. Evolution of the innate immune system: the worm perspective. Immunol Rev 2004; 198:36–58. 13. Mylonakis E, Casadevall A, Ausubel FM. Exploiting amoeboid and non-vertebrate animal model systems to study the virulence of human pathogenic fungi. PLoS Pathog 2007; 3:e101. 14. Fuchs BB, O’Brien E, Khoury JB, Mylonakis E. Methods for using Galleria mellonella as a model host to study fungal pathogenesis. Virulence 2010; 1:475–82. 15. Peleg AY, Tampakakis E, Fuchs BB, Eliopoulos GM, Moellering RC Jr, Mylonakis E. Prokaryote-eukaryote interactions identified by using Caenorhabditis elegans. Proc Natl Acad Sci U S A 2008; 105:14585–90. Echinocandins and Fungal-Bacterial Interactions • CID 2015:61 (Suppl 6) • S633 16. Tampakakis E, Peleg AY, Mylonakis E. Interaction of Candida albicans with an intestinal pathogen, Salmonella enterica serovar Typhimurium. Eukaryot Cell 2009; 8:732–7. 17. Kim Y, Mylonakis E. Killing of Candida albicans filaments by Salmonella enterica serovar Typhimurium is mediated by sopB effectors, parts of a type III secretion system. Eukaryot Cell 2011; 10:782–90. 18. Cruz MR, Graham CE, Gagliano BC, Lorenz MC, Garsin DA. Enterococcus faecalis inhibits hyphal morphogenesis and virulence of Candida albicans. Infect Immun 2013; 81:189–200. 19. Smith MG, Des Etages SG, Snyder M. Microbial synergy via an ethanoltriggered pathway. Mol Cell Biol 2004; 24:3874–84. 20. Ader F, Jawhara S, Nseir S, et al. Short term Candida albicans colonization reduces Pseudomonas aeruginosa-related lung injury and bacterial burden in a murine model. Crit Care 2011; 15:R150. 21. Roux D, Gaudry S, Khoy-Ear L, et al. Airway fungal colonization compromises the immune system allowing bacterial pneumonia to prevail. Crit Care Med 2013; 41:e191–9. 22. Tarumoto N, Kinjo Y, Kitano N, et al. Exacerbation of invasive Candida albicans infection by commensal bacteria or a glycolipid through IFN-gamma produced in part by iNKT cells. J Infect Dis 2014; 209:799–810. 23. Peters BM, Noverr MC. Candida albicans-Staphylococcus aureus polymicrobial peritonitis modulates host innate immunity. Infect Immun 2013; 81:2178–89. 24. Canadian Critical Care Trials Group. A randomized trial of diagnostic techniques for ventilator-associated pneumonia. N Engl J Med 2006; 355:2619–30. 25. Lynch AS, Robertson GT. Bacterial and fungal biofilm infections. Annu Rev Med 2008; 59:415–28. 26. Azoulay E, Timsit JF, Tafflet M, et al. Candida colonization of the respiratory tract and subsequent Pseudomonas ventilator-associated pneumonia. Chest 2006; 129:110–7. 27. Williamson DR, Albert M, Perreault MM, et al. The relationship between Candida species cultured from the respiratory tract and systemic inflammation in critically ill patients with ventilator-associated pneumonia. Can J Anaesth 2011; 58:275–84. 28. Delisle MS, Williamson DR, Albert M, et al. Impact of Candida species on clinical outcomes in patients with suspected ventilator-associated pneumonia. Can Respir J 2011; 18:131–6. 29. Delisle MS, Williamson DR, Perreault MM, Albert M, Jiang X, Heyland DK. The clinical significance of Candida colonization of respiratory tract secretions in critically ill patients. J Crit Care 2008; 23:11–7. 30. Arvanitis M, Anagnostou T, Kourkoumpetis TK, Ziakas PD, Desalermos A, Mylonakis E. The impact of antimicrobial resistance and aging in VAP outcomes: experience from a large tertiary care center. PLoS One 2014; 9:e89984. S634 • CID 2015:61 (Suppl 6) • Arvanitis and Mylonakis 31. Dyess DL, Garrison RN, Fry DE. Candida sepsis. Implications of polymicrobial blood-borne infection. Arch Surg 1985; 120:345–8. 32. Jacobs S, Price Evans DA, Tariq M, Al Omar NF. Fluconazole improves survival in septic shock: a randomized double-blind prospective study. Crit Care Med 2003; 31:1938–46. 33. Albert M, Williamson D, Muscedere J, et al. Candida in the respiratory tract secretions of critically ill patients and the impact of antifungal treatment: a randomized placebo controlled pilot trial (CANTREAT study). Intensive Care Med 2014; 40:1313–22. 34. Lazzell AL, Chaturvedi AK, Pierce CG, Prasad D, Uppuluri P, LopezRibot JL. Treatment and prevention of Candida albicans biofilms with caspofungin in a novel central venous catheter murine model of candidiasis. J Antimicrob Chemother 2009; 64:567–70. 35. Seidler M, Salvenmoser S, Muller FM. In vitro effects of micafungin against Candida biofilms on polystyrene and central venous catheter sections. Int J Antimicrob Agents 2006; 28:568–73. 36. Kucharikova S, Tournu H, Holtappels M, Van Dijck P, Lagrou K. In vivo efficacy of anidulafungin against mature Candida albicans biofilms in a novel rat model of catheter-associated candidiasis. Antimicrob Agents Chemother 2010; 54:4474–5. 37. Kucharikova S, Sharma N, Spriet I, Maertens J, Van Dijck P, Lagrou K. Activities of systemically administered echinocandins against in vivo mature Candida albicans biofilms developed in a rat subcutaneous model. Antimicrob Agents Chemother 2013; 57:2365–8. 38. Kelly J, Kavanagh K. Caspofungin primes the immune response of the larvae of Galleria mellonella and induces a non-specific antimicrobial response. J Med Microbiol 2011; 60:189–96. 39. Lamaris GA, Lewis RE, Chamilos G, et al. Caspofungin-mediated betaglucan unmasking and enhancement of human polymorphonuclear neutrophil activity against Aspergillus and non-Aspergillus hyphae. J Infect Dis 2008; 198:186–92. 40. Allizond V, Banche G, Giacchino F, et al. Candida albicans infections in renal transplant recipients: effect of caspofungin on polymorphonuclear cells. Antimicrob Agents Chemother 2011; 55:5936–8. 41. Bozza S, Gaziano R, Spreca A, et al. Dendritic cells transport conidia and hyphae of Aspergillus fumigatus from the airways to the draining lymph nodes and initiate disparate Th responses to the fungus. J Immunol 2002; 168:1362–71. 42. Ostrosky-Zeichner L, Shoham S, Vazquez J, et al. MSG-01: a randomized, double-blind, placebo-controlled trial of caspofungin prophylaxis followed by preemptive therapy for invasive candidiasis in high-risk adults in the critical care setting. Clin Infect Dis 2014; 58:1219–26. 43. Timsit JF, Azoulay E, Cornet M, et al. EMPIRICUS micafungin versus placebo during nosocomial sepsis in Candida multi-colonized ICU patients with multiple organ failures: study protocol for a randomized controlled trial. Trials 2013; 14:399.