Invasive Mycoses: Diagnostic Challenges

Transcription

1 Supplement issue Luis Ostrosky-Zeichner, MD, FACP, FIDSA Department of Medicine, University of Texas Health Medical School, Houston, Texas, USA; and Department of Epidemiology, Memorial Hermann Texas Medical Center, Houston, Texas, USA. ABSTRACT Despite the availability of newer antifungal drugs, outcomes for patients with invasive fungal infections (IFIs) continue to be poor, in large part due to delayed diagnosis and initiation of appropriate antifungal therapy. Standard histopathologic diagnostic techniques are often untenable in at-risk patients, and culture-based diagnostics typically are too insensitive or nonspecific, or provide results after too long a delay for optimal IFI management. Newer surrogate markers of IFIs with improved sensitivity and specificity are needed to enable earlier diagnosis and, ideally, to provide prognostic information and/or permit therapeutic monitoring. Surrogate assays should also be accessible and easy to implement in the hospital. Several nonculture-based assays of newer surrogates are making their way into the medical setting or are currently under investigation. These new or up-and-coming surrogates include antigens/antibodies (mannan and antimannan antibodies) or fungal metabolites (D-arabinitol) for detection of invasive candidiasis, the Aspergillus cell wall component galactomannan used to detect invasive aspergillosis, or the fungal cell wall component and panfungal marker -glucan. In addition, progress continues with use of polymerase chain reaction or other nucleic acid or molecular-based assays for diagnosis of either specific or generic IFIs, although the various methods must be better standardized before any of these approaches can be more fully implemented into the medical setting. Investigators are also beginning to explore the possibility of combining newer surrogate markers with each other or with more standard diagnostic approaches to improve sensitivity, specificity, and capacity for earlier diagnosis, at a time when fungal burden is still relatively low and more responsive to antifungal therapy Elsevier Inc. All rights reserved. The American Journal of Medicine (2012) 125, S14 S24 KEYWORDS: -Glucan; D-arabinitol; Galactomannan; Mannan; Non culture-based diagnostics; PCR assays The incidence of invasive fungal infections (IFIs) is on the rise, largely due to an increasing pool of immunocompromised or severely ill patients at elevated risk for IFIs. 1 IFIs are associated with significant morbidity and mortality, and are increasingly caused by fungal pathogens or subspecies with diminished susceptibility or resistance to many standard antifungal agents. 1 Poor outcome in patients with IFIs Statement of author disclosure: Please see the Author Disclosures section at the end of this article. This supplement is in part based on a closed roundtable meeting that was held June 7, 2011 in New York City and was jointly sponsored by Postgraduate Institute for Medicine and Global Education Exchange through an educational grant from Merck & Co., Inc. The webinar was peer-reviewed and accepted as a free multimedia activity of The American Journal of Medicine and is available at Requests for reprints should be addressed to Luis Ostrosky-Zeichner, MD, University of Texas Health Medical School, 6431 Fannin, MSB 2.112, Houston, Texas address: Luis.Ostrosky-Zeichner@uth.tmc.edu. can often be related to delayed treatment with an effective antifungal agent or combination of agents due to limitations of standard diagnostic techniques. Diagnosis of IFIs is extremely challenging, because current diagnostic methods are not sufficiently sensitive or specific, and results are often available too late to be clinically useful. 2-4 Newer diagnostic markers and techniques are available and continue to evolve, but many clinicians are unfamiliar with these approaches. Early diagnosis and/or treatment have been shown to improve patient outcomes. 5-8 Hence, there is a clear need to educate clinicians about different techniques available to diagnose and manage patients with IFIs. This article provides a general overview of the importance of early diagnosis, the need for surrogates in medical mycology, and the relative advantages and disadvantages of standard histopathologic and culture-based approaches to IFI diagnosis and newer nonculture-based diagnostic techniques /$ -see front matter 2012 Elsevier Inc. All rights reserved. doi: /j.amjmed

2 Ostrosky-Zeichner S15 CASE STUDY: PATIENT IN INTENSIVE CARE A 39-year-old Hispanic woman was admitted to the surgical intensive care unit (ICU) after a major trauma related to a motor vehicle accident. The patient was very ill, had a central line inserted, and was receiving parenteral nutrition owing to a gut injury. On day 10 of ICU admission, she presented with new-onset fever (102.2 F/39 C) and was initiated on broad-spectrum antibiotics. Blood, urine, and sputum samples were collected. After 5 days of therapy, the blood and urine culture results became available and were negative. The sputum culture was growing yeast. The lines were changed and the patient continued to be febrile. Empiric fluconazole therapy was started on day 14 of ICU admission. On day 17, the patient developed hypotension and did not respond to intravenous (IV) fluids and pressors, and subsequently died. Autopsy cultures showed Candida glabrata in her blood, liver, kidneys, and spleen. OVERVIEW OF FUNGAL DIAGNOSTICS: STANDARD TECHNIQUES Difficulties in Fungal Diagnostics The case study above illustrates the difficulties in fungal diagnostics and some of the limitations of current standard technologies. Clinical symptoms of IFIs are often nonspecific and therefore generally of little use by themselves when trying to make a timely diagnosis. Histopathologic identification of fungal pathogens in tissue samples and fungal growth using culture-based techniques are the usual means used to diagnose IFIs caused by common fungal pathogens such as Candida and Aspergillus species. 9 Unfortunately, patient populations at highest risk for IFIs are also those at high risk for complications associated with invasive biopsies, limiting the utility of histopathology. 10,11 For example, biopsy is usually not an option for patients with neutropenia with a suspected IFI such as aspergillosis, because these patients are also likely to have thrombocytopenia and be at risk for bleeding complications. 11 In addition, although microscopic examination of tissue specimens allows for rapid detection and a generic diagnosis of fungal infection, sensitivity and specificity are limited, and culturebased techniques are typically required for identification of the fungal genus or particular pathogen. 9,12,13 Blood culture is currently considered the gold standard for diagnosis of invasive candidiasis, particularly when coupled with clinical symptoms. However, blood cultures are negative for Candida in roughly 50% and 30% of patients with biopsy-proven disseminated and single-organ candidiasis, respectively, 14 meaning blood samples will miss Candida infection in 50% of patients with documented disease. In addition, it typically takes 24 to 72 hours for identification of Candida to the species level in culture. Hence, waiting for culture results before making a clinical decision means a delay in diagnosis and initiation of appropriate antifungal therapy. 9 Finally, positive cultures of nonsterile tissue specimens do not distinguish between Candida colonization versus disseminated disease, thus complicating interpretation. Diagnosis of invasive aspergillosis is even more difficult and generally is based on a constellation of patient risk factors (immune status), clinical signs/symptoms, radiologic manifestations, histologic data (when available), and microbiologic evidence, including culture results and, more recently, detection of fungal wall components like galactomannan in serum or other bodily fluids (discussed in further detail below) Negative blood cultures are the general rule for invasive aspergillosis, and hence not useful for its diagnosis, even in cases of widely disseminated disease. 16,18 By themselves, clinical signs and symptoms of invasive aspergillosis are generally vague. Chest x-rays are typically too nonspecific to be useful, and, furthermore, changes consistent with invasive pulmonary aspergillosis tend to occur late in the disease course, limiting their use for early diagnosis. 16,19 A high-resolution computed tomography (CT) scan of the chest can provide early signs consistent with pulmonary aspergillosis (e.g., the halo sign and macronodules), but these also are not specific for Aspergillus infection. 15,18-21 In addition, the halo sign appears to occur less frequently in patients without neutropenia who have pulmonary aspergillosis than in those with neutropenia, particularly when they are receiving corticosteroid therapy. 15,19,22 Histopathologic analysis is often untenable in patients with suspected aspergillosis, because (as mentioned) patients with neutropenia also tend to have thrombocytopenia, limiting use of biopsy or other invasive techniques. Even when a histopathologic analysis can be performed, typically it is not possible to distinguish Aspergillus species from other filamentous fungi, 3,15 and a failure to identify fungi in a pathology specimen does not necessarily mean absence of aspergillosis (or another IFI). 16 Cultures of bronchoalveolar lavage (BAL) fluid, sputum, or other relevant tissue or fluid compartments can be useful in the diagnosis of invasive (pulmonary) aspergillosis and identification of a particular pathogen that can be used to guide subsequent antifungal therapy, but they have the disadvantages of limited sensitivity and a relatively prolonged time for results. 3,18,23 Culture results are best interpreted in the context of risk factors, 15 because most Aspergillus culture isolates from nonsterile sites represent contamination or colonization rather than disease, 24 and the positive predictive value for invasive aspergillosis increases with rising immunosuppression. 17,24 Aspergillus species also are slow-growing fungi, meaning it may take several days to weeks for positive culture results. 3,25 In addition, culturing techniques typically require specialized expertise for recovery and species determination. 3,9 Hsu et al. 26 recently identified characteristics of an ideal fungal detection or diagnostic platform, including early detection, good sensitivity, ability to obtain species-level discrimination, detection of a broad range of fungal pathogens (multiplex capacity), reliability, quantitativity (ability to distinguish between disease and colonization), and nonin-

3 S16 The American Journal of Medicine, Vol 125, No 1A, January 2012 vasivivity. None of the standard diagnostic tests meet all these criteria, and in fact, many are lacking on multiple levels. Importance of Early Diagnosis Early diagnosis of invasive candidiasis, aspergillosis, or other IFIs is critical for optimized treatment and successful outcome. Delay in the identification of a fungal infection and specific species often enables the infection to progress to a point where subsequent antifungal therapy is ineffective or the patient dies before an appropriate regimen can be given. The case study above illustrates this point. Initial antimicrobial therapy initiated on day 10 (broad-spectrum antibiotics) and subsequent antifungal therapy initiated on day 14 (fluconazole) were not effective against the particular fungal pathogen (C glabrata), which was identified only after death. Had a diagnosis of a fungal infection involving C glabrata been available earlier, the clinicians could have initiated treatment with a more effective antifungal (e.g., an echinocandin or polyene) earlier in the disease process, presumably with a successful outcome, or at least much improved likelihood of a successful outcome. The relation between early diagnosis and treatment and mortality was illustrated in a multi-institutional study by Garey et al. 6 of 230 patients with candidemia. The results showed a significant trend for increased mortality with progressive delays in initiation of fluconazole therapy. Patient mortality rates associated with empiric therapy or zero delay (treatment initiation within 24 hours of time of blood sample collection for culture, but before organism identification) and delays of 1, 2, and 3 days were 15%, 24%, 37%, and 41%, respectively (P for trend). Morrell et al. 27 also demonstrated a relation between start of antifungal therapy and mortality in patients with candidemia, showing via multivariate regression analysis that administration of antifungal antibiotic therapy 12 hours after having the first positive blood sample for culture was a significant independent predictor of hospital mortality compared with later administration times. In the study, 5.7% of patients received treatment within 12 hours of the first positive blood sample culture for Candida species, while 6.4%, 54.8%, and 33.1% received treatment between 12 and 24 hours, between 24 and 48 hours, and 48 hours, respectively. Other studies have also reported an association between early appropriate antifungal therapy and improved outcomes in patients with either candidemia or pulmonary aspergillosis. 8,32 von Eiff and colleagues 8 observed a mortality rate of 90% for patients with pulmonary aspergillosis when antifungal therapy was initiated 10 days after onset of pneumonia. The Need for Surrogates in Medical Mycology Additional surrogate markers of fungal infection are needed in medical mycology to further improve diagnosis and treatment, and, hence, outcomes, for patients with IFIs. As discussed, standard diagnostic techniques are lacking in a number of respects, and there is a need for other surrogates with better sensitivity and specificity that can enable early diagnosis in a practical way that is easily accessible to medical laboratories and clinicians. It would also be useful if measurement of these surrogates could provide prognostic information and/or be used to monitor treatment effectiveness or patient outcomes. A number of nonculture-based diagnostic assays of surrogate markers of IFIs that offer potentially higher sensitivity and earlier diagnosis than is available with standard culture-based diagnostics are currently available or under investigation Importantly, these new markers may be the key to new strategies in medical mycology, such as prophylaxis and preemptive therapy, and improved empiric or targeted antifungal therapy. 34,37 Primary antifungal prophylaxis is used in patients without current signs or symptoms of IFI, but at high enough risk that the potential benefits of IFI prevention outweigh the risks of overuse or promotion of antifungal resistance. 38,39 These populations might include recipients of hematopoietic stem cell or solid organ transplants, patients with hematologic or solid cancers, critically ill patients, other severely immunocompromised patients, or patients undergoing major surgery or treatment involving invasive medical devices. 1,40 In this setting, surrogate markers might be used for early detection of breakthrough fungal infections, or perhaps to terminate prophylaxis when the marker indicates the patient does not have or is no longer at risk for an IFI. Newer surrogate markers might also be used to implement preemptive antifungal therapy, and perhaps to monitor the progress of such therapy. 34,41 In this situation, early detection of the surrogate in an asymptomatic patient at high risk for IFI might be interpreted as a likely IFI, and spur initiation of preemptive therapy. Preliminary results suggesting the potential use of plasma -glucan level as a surrogate marker for initiation of preemptive antifungal treatment strategy were reported by Akamatsu et al. 37 In this prospective study, a -glucan level 40 pg/ml was used as a trigger for initiation of preemptive antifungal therapy in liver transplant patients. A low mortality rate for fungal infection and mild side effects of antifungal therapy suggested the potential utility of such an approach to initiation of antifungal therapy in high-risk patients, and the authors called for further study. Initial results from other studies have suggested serum galactomannan levels might also be used as a trigger for preemptive antifungal therapy in patients with neutropenia at high-risk for IFI. 42,43 In particular, Maertens et al. 43 utilized a diagnostic and treatment algorithm incorporating the galactomannan assay (see below) and CT scans for initiation of preemptive antifungal therapy that was associated with good outcomes, together with a 78% reduction in antifungal use for neutropenic fever episodes. In addition, the protocol-driven approach would have led to early initiation of antifungal therapy in 7% of patients with invasive aspergillosis that was not clinically identified.

4 Ostrosky-Zeichner S17 At this point, these strategies should be considered experimental, but further study is clearly warranted. Empiric therapy strategies may also benefit from newer surrogate markers. Knowing that traditional markers are not very sensitive or specific, and that there is often a significant lag before results are known, these new markers may be used to make an earlier diagnosis of IFI to guide patient management. Even if the marker only enables diagnosis of a generic IFI, without identifying the specific pathogen, this can be useful information for initiation or discontinuation of empiric therapy. Likewise, if the pathogen genus, but not particular species, can be identified, this is also useful information when choosing initial empiric therapy. If the surrogate does not identify a likely IFI, this information may be used to terminate empiric antifungal therapy that has already been started, with cost benefits and reduction in risk of drug-related side effects. There are also some indications that some of the newer surrogate markers may be used to monitor ongoing therapy and to document the effectiveness or ineffectiveness of that intervention (discussed in greater detail below), and hence to modify ongoing empiric or targeted therapy. NONSTANDARD DIAGNOSTIC TECHNIQUES: UP-AND-COMING SURROGATES Newer nonculture-based assays of surrogate markers are being incorporated or investigated as means to improve the early diagnosis of IFIs caused by Candida, Aspergillus, or other fungal pathogens. Up-and-coming surrogates detected by these various assays include antibodies or antigens (e.g., mannan), metabolites (e.g., L/D-arabinitol), fungal cell wall components ( -glucan and galactomannan), and fungal-related nucleic acids, the latter utilizing molecular techniques such as polymerase chain reaction (PCR) for enhanced early detection. 33,36,44,45 Some of these assays identify multiple or essentially all clinically relevant fungi ( -glucan assays), while others are usually restricted to one or a small number of fungal pathogens (mannan, L/D-arabinitol, galactomannan assays). 36,44,46 Assays of fungal-related nucleic acids can be broad-based or limited to specific fungi, depending on the particular technique and primers utilized. 26,47 Investigators have also begun to explore whether diagnosis can be improved by a combination of methods. 36,44 Candida Diagnostics Candida-related antibody/antigen: mannan. Diagnosing Candida or other fungal infections through detection of antibodies or antigens in blood or other bodily fluids has generally been disappointing to date. Immunocompromised individuals with an IFI often produce low antibody titers, which can lead to false-negative test results, while elevated antibody levels due to Candida colonization (but not infection) can result in false-positive findings. 36 Commercially available anti-candida antibody kits (e.g., IgG Virotech [Sekisui Virotech GmbH, Rüsselsheim, Germany], IgG Biomerica [Biomerica, Inc, Irvine, CA] or SysCan3 [Rockeby Biomed Ltd., West Perth, Western Australia, Australia]). typically demonstrate either very low sensitivity or low specificity except for IgG Platelia (Bio-Rad Laboratories, Inc, Hercules, CA) which exhibits moderate sensitivity (59%) and specificity (63%) The Platelia antigen detection kit has also demonstrated reasonably high sensitivity (86%) and specificity (79%) in some studies. 48 In fact, to date, the only antibody or antigen markers that have shown any real utility for Candida diagnosis are anti-mannan antibodies and mannan antigen. Mannan is a polysaccharide and polymer of the sugar mannose that is a highly immunogenic component of the Candida cell wall. 26,51 Both mannan antigens and antibodies can increase in serum from patients with Candida infection. The Platelia Candida antigen and antibody tests are commercially available tools for the detection of serum levels of mannan antigen and antimannan antibodies, respectively, and there are some indications that detection of mannan or antimannan antibodies may be capable of playing a role in diagnosis of invasive candidiasis, if earlier results are possible than with clinical or culture-based measures. 52,53 The sensitivity and specificity of these tests have been questioned when either is used individually, but a number of reports indicate the combined sensitivity and specificity of invasive candidiasis is substantially improved when mannan antigen and antimannan antibody assays are used in combination. 53,57-59 Most recently, Mikulska and colleagues 53 reported a combined mannan/anti-mannan sensitivity and specificity for invasive candidiasis diagnosis of 83% and 86%, respectively (compared with separate sensitivities and specificities of 58% and 93% for mannan antigen alone and 59% and 83% for antimannan antibodies alone). Arendrup and associates 57 reported even higher sensitivity (100%) and somewhat lower specificity (74%) for candidemia when employing combined mannan antigen and antimannan antibody detection with lower cutoffs than those suggested by the manufacturer of the Platelia assays. The corresponding positive and negative predictive values were 80% and 100%, respectively. In summary, detection of serum mannan and antimannan antibodies is turning out to be very interesting for earlier diagnosis of invasive candidiasis, particularly when the assays are used together. This is a technique that requires close observation, as it might become a useful new addition to the diagnostic armamentarium in the next couple of years. Candida-related metabolite: D-arabinitol. D-Arabinitol is a metabolite produced by most pathogenic Candida species (e.g., Candida albicans, Candida tropicalis, Candida parapsilosis, Candida pseudotropicalis), but notably not C glabrata or Candida krusei. 26,36,60 Serum levels of D- arabinitol have been demonstrated to increase in patients with invasive candidiasis (although not necessarily those involving C glabrata or C krusei 61 ), and to distinguish individuals with invasive candidiasis from healthy controls or individuals with Candida colonization However, D-arabinitol levels may also be elevated due to abnormal

5 S18 The American Journal of Medicine, Vol 125, No 1A, January 2012 renal function. 36,66,68,69 Because of this, and because the rate of D-arabinitol clearance in urine is quantitatively equivalent to the rate of creatinine clearance, 70 investigators or clinicians often use the serum or urinary D-arabinitol/ creatinine ratio to control for renal function. 36,61,63,67,71 Importantly, assays of serum D-arabinitol have been reported to enable earlier diagnosis of invasive candidiasis than microbiologic cultures, whether measuring serum D-arabinitol itself 72 or the D-arabinitol/creatinine ratio. 61,71 Investigators have also argued that the diagnostic sensitivity and specificity of D-arabinitol may be improved by measurement of serum or urinary D-arabinitol/L-arabinitol ratios. 36,73 Some metabolic pathways in humans (and other vertebrates) produce L-arabinitol, but the principal source of D-arabinitol in human serum or urine is generated via fungal metabolism. Furthermore, serum/urinary levels of L-arabinitol are not elevated above normal (and may actually be decreased) in patients with invasive candidiasis, 36,62,74 and while abnormal renal function can produce some increase in D-arabinitol levels, the D-arabinitol/L-arabinitol ratio does not appear to significantly change. 75 In an early study by Lehtonen et al., 73 an assay measuring urine D-arabinitol/Larabinitol ratio exhibited 88% sensitivity and 91% specificity for detection of disseminated candidiasis. It was able to significantly discriminate patients with disseminated candidiasis from patients colonized with hematologia, controls with neutropenia but without invasive yeast infection, or healthy adults (P between patients with disseminated candidiasis and all control groups). However, although measurement of the D-arabinitol/Larabinitol ratio or D-arabinitol/creatinine ratio are interesting, and there are some reports of early diagnosis with relatively high sensitivity and specificity, a variety of assays exist that are not easily implemented in the clinical laboratory at this time. In addition, some investigations have suggested it is currently better suited for surveillance testing using serial measurements than for diagnosis based on a single value. 26 Further studies are needed to better characterize the potential roles of D-arabinitol assays for diagnosis or surveillance of invasive candidiasis. Aspergillus Diagnostics: Galactomannan Galactomannan is a component of the Aspergillus species cell wall that is released into the surrounding environment during fungal growth or tissue invasion. 16,26 Serum levels of galactomannan correlate with Aspergillus fungal burden in animals with experimental pulmonary aspergillosis, 76,77 and can be used to diagnose invasive aspergillosis in humans. In 2003, the United States Food and Drug Administration (FDA) approved the use of a double-sandwich enzyme immunoassay (EIA) of galactomannan (Platelia Aspergillus EIA; Bio-Rad Laboratories, Inc.) for the diagnosis of invasive aspergillosis in adult patients. The assay kit has also been available in Europe for 10 years. 78 Assay results are reported in terms of the galactomannan index (GMI) or optical density index (ODI), which represents the ratio of optical density of the patient sample compared with the optical density for a threshold serum provided in the assay kit. 26 A GMI at or above a certain threshold is interpreted as evidence of invasive aspergillosis. A GMI of 0.5 is the cutoff for positivity currently approved by the FDA. Use of a higher cutoff value increases the likelihood of false-negative results, while a lower cutoff increases sensitivity and the risk of false-positive results. The Platelia galactomannan assay has a number of attractive features. It is well standardized and user friendly. It only requires 300 ml for analysis and has a fast turnaround time ( 4 hours). In addition, the assay kit can be divided into several parts, so the user does not have to use up the entire kit if only a few samples are being tested. Given that it is an EIA-based technique, it should be available in most well-equipped microbiology laboratories. Sensitivity and specificity are generally high, 79 and, as mentioned, the assay is approved in both the United States and Europe. In addition, the 2008 Infectious Diseases Society of America (IDSA) clinical practice guidelines for aspergillosis endorse the assay as a useful adjunctive test to establish an early diagnosis, particularly when used in serial screening of patients at high risk of infection. 80 Lastly, some data suggest serial measurements of galactomannan may be useful for monitoring therapeutic progress However, as the IDSA guidelines make clear, use of galactomannan measurements for therapeutic monitoring in patients with invasive aspergillosis should only be considered experimental at this time. 80 However, there are some potential problems with using the galactomannan assays for diagnosis of invasive aspergillosis. Although the test is generally associated with high sensitivity and specificity for invasive aspergillosis, some variability has been observed in these values and its predictive value, 16,26 perhaps owing in part to the several variables that have been shown to affect assay results. In particular, false-positive results can occur due to use of -lactam antibiotics (e.g., piperacillin-tazobactam, amoxicillin-clavulanate), presence of certain other invasive mycoses (e.g., Penicillium, histoplasmosis, or blastomycosis), or use of plasmalyte in BAL fluid or other infusion solutions. 26,80 False-negative results can be owing to prior or concomitant antifungal therapy, an infection that has been walled off, or low fungal burden. 16,26 Sensitivity of the assay for invasive aspergillosis also appears to be lower in patients with solid organ transplant than hematopoietic stem cell transplant recipients or those with a hematologic malignancy. 44,79 The galactomannan assay was initially studied and approved for use in the evaluation of serum samples for diagnosis of invasive aspergillosis. However, the galactomannan assay can be used with other specimens, and has increasingly been used for measurement of galactomannan in BAL fluid in patients suspected of pulmonary aspergillosis A 2004 review by Klont and colleagues 88 of assay-detected galactomannan in nonserum samples reported variable performance in BAL and urine samples, excellent performance from limited analyses of cerebrospinal fluid, and only anecdotal results for use of the assay in

6 Ostrosky-Zeichner S19 cyst fluid/pus or direct tissue. With respect to BAL galactomannan, some research has reported very high sensitivity, while other studies a high rate of false-positive findings, particularly in patients with airways colonized with Aspergillus. Given the controversy in the area, the galactomannan assay should be used with caution for BAL fluid collected from patients suspected of invasive pulmonary aspergillosis, being mindful of the potential for a false-positive result owing to airway colonization. -Glucan as a Panfungal Marker Unlike galactomannan, which is found selectively in Aspergillus species, -glucan is a polysaccharide located in the cell membrane of most fungal pathogens, notable exceptions being Mucor species (formerly known as zygomycetes) and Cryptococcus species, which either do not express -glucan or exhibit only low-level or variable expression. 4,26,36 Because -glucan is expressed in most fungi, but not in mammals, bacteria, or viruses, 36 detection of -glucan in blood or other bodily specimens represents a reasonable (essentially) panfungal marker for IFI. The Fungitell Assay (Associates of Cape Cod, Inc., East Falmouth, MA) is a commercially available and FDA-approved assay for the detection of -glucan in serum, useful for IFI diagnosis. 26,89 Additional -glucan assays are available in other countries, but are not approved for use in the United States. 26,36 Small quantities of -glucan are released into the blood during an IFI, and the Fungitell Assay can detect -glucan in serum and provide results within 2 hours. 89 The assay utilizes a blood coagulation factor (factor G) isolated from blood cells of the North American Horseshoe Crab to selectively detect -glucan using a colorimetric assay. 26,89 -Glucan detection has been favorably discussed in guidelines from the IDSA for the diagnosis of candidiasis 90 or aspergillosis, 80 and in the European Organization for Research and Treatment (EORTC)/Mycosis Study Group (MSG) of the National Institute of Allergy and Infectious Diseases (NIAID) revised definitions of IFI as a marker for probable IFI. 91 An important study by a group of Japanese researchers appeared in Lancet in 1995, demonstrating high sensitivity (90%), specificity (100%), and negative predictive value (97%) for invasive mycosis with use of a -glucan assay (similar but not identical to Fungitell) and a plasma cutoff value of 20 pg/ml in febrile patients in the hospital. 46 With a few exceptions, subsequent studies performed in the United States and elsewhere have generally confirmed the high specificity and sensitivity (reviewed by Wheat 92 and Karageorgopoulos et al. 93 ) and high negative predictive value of -glucan assays when a suitable cutoff value is used as a marker for candidiasis, aspergillosis, or generic IFI. A recent meta-analysis of 16 studies measuring serum or plasma -glucan for the diagnosis of IFIs (2979 total patients, 594 with proven or probable IFIs) reported pooled sensitivity and specificity values of 77% and 85%, respectively. 93 The area-under-the-summary hierarchical receiver operating characteristic curve of the sensitivity versus specificity of measuring -glucan levels for diagnosis of proven/ probable IFI was 0.89, where a value between 0.80 to 0.90 is generally considered representative of good diagnostic test accuracy. There was no notable difference in the sensitivity of -glucan testing for detection of invasive candidiasis or invasive aspergillosis. In summary, advantages of -glucan testing for diagnosis include its ability to detect a wide range of IFIs (panfungal marker) and its generally high sensitivity and especially specificity. Because it has a strong negative predictive value, -glucan testing can be used to exclude an IFI as well as to diagnose one. In addition, the performance of -glucan testing does not appear to be affected by prior antifungal therapy. 97 Because there is a commercially available kit in the United States, the assay can be routinely used in most hospitals in this country. However, there are a number of potential drawbacks associated with -glucan testing for IFI diagnosis. First, although the Fungitell assay is readily available and can be routinely used in the United States, it is not necessarily user-friendly. Because of the assay format, investigators must use the whole plate when performing the test, even if only a few samples are being run. This means most institutions end up sending the samples to a reference laboratory, which increases turnaround time and removes a potential benefit of the test, and discourages many clinicians from routinely using the test. In addition, although the ability of -glucan testing to detect a wide range of fungi can be considered a positive feature, this can also be a negative feature when more specific information about the particular pathogenic fungal is desired. Furthermore, a number of factors can also lead to false-positive results, including administration of immunoglobulin or albumin preparations contaminated with fungal components, hemodialysis with cellulose membranes, presence of serious bacterial infection, packing of serosal surfaces with gauze containing glucan, and possibly use of certain antibiotics (e.g., piperacillin-tazobactam). 4,26,98 Lastly, as mentioned, -glucan testing cannot be used to reliably detect mucormycosis or cryptococcosis. Nonetheless, after weighing the pros and cons, there is enough to support more routine use of -glucan testing at this time. Initial studies suggest that -glucan assays may be combined with either measurement of galactomannan 99 or detection of antibodies to Candida-specific antigens 100 to improve sensitivity and positive predictive value. Further studies in this area are warranted. PCR-Based Diagnostics Nucleic acid-based diagnostic techniques are perhaps the fastest growing segment of fungal diagnostics in many clinical laboratories. 101 Generally speaking, molecular-based diagnostics can detect IFIs with high sensitivity and may provide results more rapidly than standard diagnostic approaches, thereby enabling the possibility of earlier diagnosis and more timely initiation of antifungal therapy, when appropriate. 4,9,23,36,101 In addition, certain molecular-based

7 S20 The American Journal of Medicine, Vol 125, No 1A, January 2012 Table 1 Diagnostic Candida Polymerase Chain Reaction Meta-Analyses (hierarchical summary receiver operating characteristic curves) a Likelihood Ratio Comparison with TN at-risk patients (no. of studies) Sensitivity Specificity DOR Positive Negative TP I patients (with candidemia) All b (49) 0.96 ( ) 0.92 ( ) ( ) 12.3 ( ) 0.05 ( ) Year, 2000 (27) 0.93 ( ) 0.94 ( ) ( ) 15.4 (9-26.4) 0.08 ( ) TP II patients (with proven/probable IC) All b (17) 0.93 ( ) 0.95 ( ) ( ) 18.2 (7-47.2) 0.07 ( ) Year, 2000 (8) 0.87 ( ) 0.96 ( ) ( ) 24.3 ( ) 0.14 ( ) TP III patients (with proven/probable/ possible IC) All b (20) 0.73 ( ) 0.91 ( ) 26.9 ( ) 8.1 (4-16.5) 0.3 ( ) Year, 2000 (11) 0.71 ( ) 0.95 ( ) 47.3 ( ) 14.4 (8-25.2) 0.3 ( ) TN true-negative; TP true-positive; DOR diagnostic odds ratio; IC invasive candidiasis. Adapted from Avni et al, a Values in parentheses are 95% confidence intervals. b Includes data from all studies published between 1993 and 2009 that met the inclusion criteria for the systematic review and subgroup analysis. techniques can establish a diagnosis at the species level or be used for rapid assessment of primary or secondary resistance. 23, Many molecular platforms are currently under investigation as diagnostic tools. These varied approaches use a wide range of primers and exhibit great variation in a number of other technical aspects related to methodology, which can lead to a range of reported sensitivities and specificities. 4,16,26,105 In fact, it is largely lack of standardization of methods and varied results across laboratories that has prevented nucleic acid based diagnostics from fulfilling its promise and becoming a more widely utilized diagnostic approach in the current environment. Other potential disadvantages of nucleic acid based diagnostics include difficulty in distinguishing fungal colonization from disease and potential for false-positive results owing to contamination or other causes. 23,26,36 Because of these shortcomings, nucleic acid based diagnostic techniques must be considered investigational at this time. There are currently no FDA-approved nucleic acid based assays for fungal diagnosis. PCR-based assays are currently the most commonly employed nucleic acid or molecular-based approach to fungal diagnostics, 4 and real-time PCR (RT-PCR) has been a particularly great advancement in the field, 9,36 although the problems of lack of standardization pertain to RT-PCR and other PCR-based assays as they do for other nucleic acid based approaches. Recent studies have reported reliable fungal pathogen identification and characterization in 3 hours with RT-PCR. 106,107 Most PCR- and other nucleic acid based diagnostic assays studied to date have either focused on detection of Candida or Aspergillus species, or have used primers enabling panfungal detection. 4,26,36 A recent meta-analysis of 54 studies using PCR-based methods to diagnose invasive candidiasis from blood samples concluded that the use of direct PCR is associated with good sensitivity and specificity for rapid diagnosis of invasive candidiasis. 108 However, PCR performance was observed to correlate with the degree of certainty of the diagnosis, as illustrated in Table Sensitivity and specificity are very good when the sample includes only patients with proven candidemia, but sensitivity tends to decrease when individuals with probable or possible invasive candidiasis are added to the population. Based on this, one could conclude that investigators must standardize the way these markers are studied using the standard definitions of invasive candidiasis, such as those from the consensus group of the EORTC and MSG. 91,109 A study by Cuenca-Estrella and colleagues 110 suggests that combining serial RT-PCR for invasive aspergillosis with galactomannan testing improves sensitivity and early diagnosis of invasive aspergillosis. Additional studies of PCR-based technologies in combination with other diagnostic modalities should better clarify the feasibility and potential benefits of combination approaches to diagnosis of invasive candidiasis, aspergillosis, or other IFIs. CASE STUDY: A PATIENT WITH LEUKEMIA A 25-year-old white woman was recently diagnosed with acute myeloid leukemia (AML). She was initiated on chemotherapy and began undergoing -glucan surveillance. On day 7 after induction chemotherapy, the patient s absolute neutrophil count was /L. At this time, the -glucan assay gave a value of 1,073 pg/ml of blood. About 2 days later, she developed a persistent fever. The patient was initiated on liposomal amphotericin B therapy, and her temperature began to decrease, along with the -glucan level. The -glucan level remained low for a number of days before starting to increase again sometime after day 10 post induction chemotherapy. The patient eventually died on day 18, and the autopsy revealed disseminated fusariosis. Figure 1

8 Ostrosky-Zeichner S21 Figure 1 -Glucan levels during the course of chemotherapy and antifungal therapy in a patient with acute myeloid leukemia. BDG -D-glucan. represents the findings from time of initiation of -glucan surveillance to the patient s death 18 days later. Analysis Several points can be made based on this case study. First, a positive serum -glucan 1,000 pg/ml is high and consistent with an IFI. A clinician should be comfortable making a diagnosis of IFI based on this information, if there are no other potential reasons for the high -glucan, e.g., use of -lactams or recent surgery or hemodialysis. The subsequent declines in both fever and -glucan level after initiation of antifungal therapy with liposomal amphotericin B further support the initial diagnosis of IFI, and suggest it is caused by a pathogen susceptible to liposomal amphotericin B therapy. The concomitant declines in fever and -glucan, together with the subsequent increase in -glucan level as the patient s condition declines, also suggest that changes in serum levels of -glucan correspond with outcome, and that serial measurements of serum -glucan level may be useful in evaluating treatment course in a given patient over time. The changes in -glucan level and patient health over time for this case also may be explained by initial infection with Candida or another fungus receptive to liposomal amphotericin B, followed by a breakthrough infection involving a pathogen unresponsive to liposomal amphotericin B. This, of course, cannot be conclusively determined by use of a single marker. Nonetheless, the data reviewed here indicate the utility of -glucan for the diagnosis of (generic) IFI and its possible use to evaluate treatment effectiveness over time. SUMMARY Diagnosis of fungal infections can be challenging. Not only are better diagnostic techniques or strategies needed for accurate diagnosis, but there also is a great need for approaches enabling an early diagnosis, at a time when the IFI is less advanced and more susceptible to appropriate antifungal therapy. Given the changing epidemiology of IFIs and increasing involvement of genera and species with reduced susceptibility or resistance to several conventional antifungal agents, early identification of the particular pathogenic species is important. Several studies have demonstrated that early diagnosis and antifungal therapy are associated with better outcomes, and specifically lowered mortality rates in patients with either candidemia or aspergillosis. Traditional diagnostics are slow and have limited sensitivity and/or specificity. By comparison, new diagnostics such as galactomannan, -glucan, and PCR assays offer opportunities for newer early treatment strategies, such as preemptive therapy, with improved outcomes. They may also offer opportunities for more rational antifungal use, including discontinuation and/or modification of initial empiric antimicrobial therapy. Clinicians must be aware of and familiar with how best to utilize these newer diagnostics, in conjunction with older diagnostic techniques, to improve the care of patients with a proven, probable, or suspected IFI. ACKNOWLEDGMENT Editorial support for this publication was provided by Global Education Exchange, Inc., Freehold, New Jersey.

9 S22 The American Journal of Medicine, Vol 125, No 1A, January 2012 AUTHOR DISCLOSURES The author of this article has disclosed the following industry relationships: Luis Ostrosky-Zeichner, MD, has received consulting and/or non-cme service fees from Astellas, Glaxo- SmithKline, Merck & Co., Inc., NovaDigm Therapeutics, Pfizer and grants for contracted research from Associates of Cape Cod, Inc., Astellas, Merck & Co., Inc., and Pfizer. References 1. Pfaller MA, Diekema DJ. Epidemiology of invasive mycoses in North America. Crit Rev Microbiol. 2010;36: Ellepola AN, Morrison CJ. Laboratory diagnosis of invasive candidiasis. J Microbiol. 2005;43(spec no): Hope WW, Walsh TJ, Denning DW. Laboratory diagnosis of invasive aspergillosis. Lancet Infect Dis. 2005;5: Preuner S, Lion T. Towards molecular diagnostics of invasive fungal infections. Expert Rev Mol Diagn. 2009;9: Bassetti M, Trecarichi EM, Righi E, et al. Incidence, risk factors, and predictors of outcome of candidemia: survey in 2 Italian university hospitals. Diagn Microbiol Infect Dis. 2007;58: Garey KW, Rege M, Pai MP, et al. Time to initiation of fluconazole therapy impacts mortality in patients with candidemia: a multi-institutional study. Clin Infect Dis. 2006;43: Golan Y, Wolf MP, Pauker SG, Wong JB, Hadley S. Empirical anti-candida therapy among selected patients in the intensive care unit: a cost-effectiveness analysis. Ann Intern Med. 2005;143: von Eiff M, Roos N, Schulten R, Hesse M, Zuhlsdorf M, van de Loo J. Pulmonary aspergillosis: early diagnosis improves survival. Respiration. 1995;62: Wengenack NL, Binnicker MJ. Fungal molecular diagnostics. Clin Chest Med. 2009;30: , viii. 10. Alexander BD. Diagnosis of fungal infection: new technologies for the mycology laboratory. Transpl Infect Dis. 2002;4(suppl 3): Vehreschild JJ, Ruping MJ, Steinbach A, Cornely OA. Diagnosis and treatment of fungal infections in allogeneic stem cell and solid organ transplant recipients. Expert Opin Pharmacother. 2010;11: Montagna MT, Caggiano G, Borghi E, Morace G. The role of the laboratory in the diagnosis of invasive candidiasis. Drugs. 2009; 69(suppl 1): Peman J, Zaragoza R. Current diagnostic approaches to invasive candidiasis in critical care settings. Mycoses. 2010;53: Berenguer J, Buck M, Witebsky F, Stock F, Pizzo PA, Walsh TJ. Lysis-centrifugation blood cultures in the detection of tissue-proven invasive candidiasis: disseminated versus single-organ infection. Diagn Microbiol Infect Dis. 1993;17: Barnes PD, Marr KA. Aspergillosis: spectrum of disease, diagnosis, and treatment. Infect Dis Clin North Am. 2006;20: , vi. 16. Barnes PD, Marr KA. Risks, diagnosis and outcomes of invasive fungal infections in haematopoietic stem cell transplant recipients. Br J Haematol. 2007;139: Horvath JA, Dummer S. The use of respiratory-tract cultures in the diagnosis of invasive pulmonary aspergillosis. Am J Med. 1996;100: Sales Mda P. Chapter 5 Aspergillosis: from diagnosis to treatment. J Bras Pneumol. 2009;35: Barnes RA. Early diagnosis of fungal infection in immunocompromised patients. J Antimicrob Chemother. 2008;61(suppl 1):i3-i Georgiadou SP, Sipsas NV, Marom EM, Kontoyiannis DP. The diagnostic value of halo and reversed halo signs for invasive mold infections in compromised hosts. Clin Infect Dis. 2011;52: Gotway MB, Dawn SK, Caoili EM, Reddy GP, Araoz PA, Webb WR. The radiologic spectrum of pulmonary Aspergillus infections. J Comput Assist Tomogr. 2002;26: Cornillet A, Camus C, Nimubona S, et al. Comparison of epidemiological, clinical, and biological features of invasive aspergillosis in neutropenic and nonneutropenic patients: a 6-year survey. Clin Infect Dis. 2006;43: Sherif R, Segal BH. Pulmonary aspergillosis: clinical presentation, diagnostic tests, management and complications. Curr Opin Pulm Med. 2010;16: Perfect JR, Cox GM, Lee JY, et al. The impact of culture isolation of Aspergillus species: a hospital-based survey of aspergillosis. Clin Infect Dis. 2001;33: Duthie R, Denning DW. Aspergillus fungemia: report of two cases and review. Clin Infect Dis. 1995;20: Hsu JL, Ruoss SJ, Bower ND, Lin M, Holodniy M, Stevens DA. Diagnosing invasive fungal disease in critically ill patients. Crit Rev Microbiol. 2011;37: Morrell M, Fraser VJ, Kollef MH. Delaying the empiric treatment of Candida bloodstream infection until positive blood culture results are obtained: a potential risk factor for hospital mortality. Antimicrob Agents Chemother. 2005;49: Nolla-Salas J, Sitges-Serra A, Leon-Gil C, et al, for the Study Group of Fungal Infection in the ICU. Candidemia in non-neutropenic critically ill patients: analysis of prognostic factors and assessment of systemic antifungal therapy. Intensive Care Med. 1997;23: Patel GP, Simon D, Scheetz M, Crank CW, Lodise T, Patel N. The effect of time to antifungal therapy on mortality in candidemiaassociated septic shock. Am J Ther. 2009;16: Tumbarello M, Posteraro B, Trecarichi EM, et al. Biofilm production by Candida species and inadequate antifungal therapy as predictors of mortality for patients with candidemia. J Clin Microbiol. 2007;45: Zilberberg MD, Kollef MH, Arnold H, et al. Inappropriate empiric antifungal therapy for candidemia in the ICU and hospital resource utilization: a retrospective cohort study. BMC Infect Dis. 2010;10: Caillot D, Casasnovas O, Bernard A, et al. Improved management of invasive pulmonary aspergillosis in neutropenic patients using early thoracic computed tomographic scan and surgery. J Clin Oncol. 1997;15: Chen SC, Kontoyiannis DP. New molecular and surrogate biomarkerbased tests in the diagnosis of bacterial and fungal infection in febrile neutropenic patients. Curr Opin Infect Dis. 2010;23: Einsele H, Loeffler J. Contribution of new diagnostic approaches to antifungal treatment plans in high-risk haematology patients. Clin Microbiol Infect. 2008;14(suppl 4): Guery BP, Arendrup MC, Auzinger G, et al. Management of invasive candidiasis and candidemia in adult non-neutropenic intensive care unit patients: part I. Epidemiology and diagnosis. Intensive Care Med. 2009;35: Lain A, Elguezabal N, Moragues MD, Garcia-Ruiz JC, del Palacio A, Ponton J. Contribution of serum biomarkers to the diagnosis of invasive candidiasis. Expert Rev Mol Diagn. 2008;8: Akamatsu N, Sugawara Y, Kaneko J, Tamura S, Makuuchi M. Preemptive treatment of fungal infection based on plasma (1 3) -D-glucan levels after liver transplantation. Infection. 2007;35: Mikolajewska A, Schwartz S, Ruhnke M. Antifungal treatment strategies in patients with haematological diseases or cancer: from prophylaxis to empirical, pre-emptive and targeted therapy. Mycoses. 2011:Epub ahead of print. 39. Ruping MJ, Vehreschild JJ, Cornely OA. Patients at high risk of invasive fungal infections: when and how to treat. Drugs. 2008;68: Morace G, Borghi E. Fungal infections in ICU patients: epidemiology and the role of diagnostics. Minerva Anestesiol. 2010;76: