- Email: [email protected]
Invasive fungal infections: evolving challenges for diagnosis and therapeutics
Molecular Immunology 38 (2001) 947–957
Invasive fungal infections: evolving challenges for diagnosis and therapeutics Michael Ellis∗ Faculty of Medicine and Health Sciences and Tawam Hospital, Al Ain, Abu Dhabi, United Arab Emirates Received 15 October 2001; accepted 10 December 2001
Abstract Invasive fungal infections (IFI) parallel the explosive increase in the immunocompromized patient population, and are characterized by diagnostic difficulties and extreme mortality. Candidemia in a tertiary referral hospital in the Middle East confirms the current epidemiologic shift in this common blood stream pathogen towards non-malignancy cases (38%) and antifungal prophylaxis failure (20%), high presentation sepsis scores and attributable mortality (32%). Invasive aspergillosis (IA) is also associated with high mortality. Use of non-invasive computerized tomographic (CT) radiologic scanning linked to early administration of high dose liposomal amphotericin B (LAB) is associated with a reduced mortality of 9.5% compared to historical experience of 28%. Life threatening invasive aspergillosis also occurs in patients who are less obviously immunocompromized. Investigations may reveal subtle immune deficits which could place the patient at some risk for an invasive mycosis. Antifungal treatment used in combination with progenitor cell growth factors and ␥-interferon has proved successful in such situations of progressive fungal disease unresponsive to antifungal therapy alone. Pharmacologic remodeling of existing compounds by lipidisation reduces both the toxicity denominator and the efficacy numerator of the therapeutic index when compared to the parent drug. A comparative dose study of liposomal amphotericin B in aspergillosis has demonstrated equi-efficacy, generated debate over the ability of the controlled clinical trial to be capable of assessing antifungal efficacy, and illustrated that recovery from an invasive fungal infection may require maximum tolerated doses and immunomanipulation. Several new antifungal strategies are under clinical investigation. These include reformulating existing antifungals, exploitation of the growing knowledge of virulence factors to synthesize antagonists, immune reconstitution and immunoprotection. An interim analysis of an ongoing placebo controlled study of recombinant interleukin-11 to assess its efficacy in reducing sepsis in leukemia patients through prevention of chemotherapy induced gut epithelial cell apoptosis, has demonstrated a difference in the two study arms in sepsis rates and preservation of gastrointestinal epithelial cell integrity. The unique and special challenges presented by the dynamic epidemiologics of invasive fungal infections are demanding and attracting considerable responses, in the fields of diagnosis and therapeutics. Current strategies need considerable improvement, yet ongoing collaborative efforts will have a positive impact on our understanding of the fungus–host interaction and ultimately our ability to offer better care for our patients with invasive mycoses. © 2002 Published by Elsevier Science Ltd. Keywords: Candida; Aspergillus; Amphotericin B; Immunocompromization; Halo sign; Cytokines
1. Introduction
1.1. The scope of the problem
This article focuses on the current status of invasive fungal infections (IFI) from a clinical perspective. It reviews the major challenges in diagnostics and treatments and includes local case studies and preliminary results of clinical studies conducted at Tawam Hospital, UAE.
The end of the 20th century witnessed a remarkable change in the pattern of pathogens responsible for serious human infections. Among hospitalized patients fungal etiology of pneumonia, urinary tract infection, surgical wound infection and blood stream infections had increased 50, 100, 200 and 400%, respectively from 1990 through 2000 (Jarvis, 1995). Candidemia had risen to rank 4th among blood stream pathogens in the US (Abi-Said et al., 1997), its
∗ Fax: +971-3-722-3942. E-mail address: [email protected] (M. Ellis).
0161-5890/02/$ – see front matter © 2002 Published by Elsevier Science Ltd. PII: S 0 1 6 1 - 5 8 9 0 ( 0 2 ) 0 0 0 2 2 - 6
948
M. Ellis / Molecular Immunology 38 (2001) 947–957
lethality reflected in the high crude mortality rate of 40%. A unique population based surveillance study in 1998 from San Francisco gave good evidence for the first time that IFI were not just a phenomenon confined to more severely immunocompromized hospital patients, but that the problem was endemic throughout the general population (Rees et al., 1998). One of the surprising findings from that study was the exponential explosive rise of invasive candidiasis with age—an indicator also that IFI do not occur solely in the more seriously or classically immunocompromized.
2. Epidemiology and causality A heterogeneity of factors is responsible for the emergence of invasive mycoses as a significant cause of morbidity and mortality. These include: the increased bipolarity of age at the one end characterized by a striking increase in pre-term births (16% increase in 10 years in some US population) and at the other by the significant increased pool of the “super-old”, socio-behavioral disintegration associated with HIV, increased survivorship of people with congenital immunodeficiency syndromes, rigorous and repetitive abdominal surgery, widespread use and abuse of antimicrobials, intrusive vascular lines, transmucosal devices and total parental nutrition support systems. It is simply not the disproportionate increase in IFI that is occurring but also a notable epidemiologic shift in the types of molds and yeasts. The relative contribution of Aspergillus, as shown in autopsy studies has risen steadily, replacing Candida as the top pathogen in some institutions (Groll et al., 1996). Fusarium was once confined to an uninteresting dermatophyte role—now it has moved out of its ecologically secure niche to become second to Aspergillus as a major invasive mold, mimicking aspergillosis clinically and radiologically. Penicillium spp. are in general innocuous, but now life threatening TB-like infections are described; penicilliosis has became the 3rd commonest AIDS defining infection in Thailand. In 1963 only five Candida spp. were known to cause human disease—now more than 20 yeasts are commonly reported including Malassezia, Rhodotorula and Prototheca spp. Among the invasive Candida spp. the non-albicans species have emerged, accounting for up to 60% of all the candidemias seen in some hospitals (Pfaller et al., 1998a,b)—the most recent being Candida ciferrei with marked azole resistance (Gunsilius et al., 2001)—bringing challenges for effective antifungal therapy. This epidemiologic shift is not clearly understood. The widespread use of azoles has been implicated (Viscoli et al., 1999) but a causal relation remains unproven. Retrospective and uncontrolled data analysis, single center sources (institution bias), nosocomial outbreaks (substantiated by molecular epidemiologics), increase in non-albicans which pre-date azole use and greatly increased chemotherapeutic regimens
Fig. 1. Proportion of C. albicans: non-albicans isolated from surveillance swabs during a 3-week period of fluconazole prophylaxis.
are important confounding variables which could also contribute to this shift. Furthermore the epidemiologic data of C. krusei (an intrinsically azole resistant species) as a surrogate marker for the azole drive to emergence of resistant non-albicans species, does not support this supposition. Thus there is a constant, rather than a rising, incidence of this species over many years at oncology centers and other large populations (e.g. HIV infected persons) in which azole use is high (Powderley et al., 1995). Nevertheless some prospective analyses do show that superficial fungal colonization (in contrast to invasive candidiasis) is subject to azole pressure. During 3 weeks of fluconazole prophylaxis, the proportion of non-albicans Candida spp. isolated from throat, urine and stool surveillance cultures substantially and significantly increased (Ellis et al., 1994; Fig. 1). Yet the invasive non-albicans Candida isolates occurred only in those patients not receiving fluconazole prophylaxis. These observations suggest other factors additional (or in spite of) azole use influence the current shift towards non-albicans candidiasis. The local experience at Tawam Hospital reflects the hospital immunocompromized scene with regard to candidiasis. Over a 5-year period there have been 60 episodes of candidemia. The expected association with hematologic malignancy is seen but the shift towards non-malignant conditions such as repetitive abdominal surgery and prematurity is striking (Fig. 2a). The other well recognized risk factors are confirmed but the high association with recent bacteremias is a novel finding. Breakthrough candidemia which should have been prevented by effective antifungal prophylaxis occurred in 20% of the patients and is a cause for concern. Over half of the isolates were non-albicans spp. (Fig. 2b). High sepsis scores and multi-organ failure dominated the clinical presentation (Table 1A). The mortality rate of 50%
M. Ellis / Molecular Immunology 38 (2001) 947–957
949
3. Current diagnostic tools Classic blood culture proves positive in less than 5% of patients in invasive aspergillosis (IA) and in 50% in invasive candidiasis. Invasive techniques such as open lung biopsy currently provide the greatest and most accurate diagnostic yield. In a unique prospective study in which two invasive diagnostic techniques were concurrently performed the yield of specific diagnosis made by open lung biopsy was three-fold greater than bronchoalveolar lavage (Ellis et al., 1995). It is important to note that the range of diagnoses is wide and infections contribute only a proportion of the etiologies in neutropenic fever patients. However, such invasive techniques to prove an invasive mycosis at the point of target organ impact is very often impossible to undertake because of the critical coagulopathic clinical profile of these patients. Currently, the diagnostic approach centers on the following three areas. 3.1. High diagnostic acumen
Fig. 2. Candidemia at Tawam Hospital: (a) underlying disease categories; (b) Candida species isolated.
gives credence to the current clinical practice guidelines that ALL candidemias should be energetically managed. Of all variables tested, shock, organ failure, persistent neutropenia and use of growth factors were independent factors significantly predictive of death (Table 1B). The patient with leukemia and lymphoma is the most immunologically disprivileged since the neutrophil which forms the pivotal host defence has been removed and its function impaired through disease and chemotherapy. The fact that 25% of patients with hematologic malignancy die with IFI—many undiagnosed in life—underscores our inability to diagnose and manage effectively.
In a profound protracted neutropenic febrile patient, with no bacterial sepsis or any other cause for fever and no clinical focus, who is failing to respond to broad spectrum antibiotics (antibiotic resistant neutropenic fever or ARNF), IA or other IFI has to be strongly suspected (EORTC, 1989). Currently serologic diagnostics are increasingly available, but these have to be validated thoroughly in the clinical setting. In addition they are expensive to use for daily testing. The bedside clinical science therefore remains of paramount importance. This was underscored by Gerson et al. (1985) who combined 11 clinical parameters into a discriminant scorecard to distinguish patients with IA from controls. Retrospective and prospective phased trials gave sensitivities of 63–93% and specificities of 88–98%. These figures are virtually identical to those currently reported for Aspergillus galactomannan testing (Verweij et al., 1996). The diagnosis of IA using the clinical scorecard was made 4 days earlier compare to a more classical clinical approach. Patients with the untreated ARNF syndrome are at a high risk for rapid overt IFI.
Table 1 Candidemia (A) Characteristics at presentation (Tawam Hospital)
(B) Variables associated with death
Characteristics
Percentage
Variable
P
Shock Renal failure Multi-organ failure Sepsis score ≥ 3 Mortality crude Attributable Associated
43 23 33 68 50 32 13
ICU admission Malignant disease non-remission Shock at presentation Renal failure at presentation Multi-organ failure at presentation Non-removal central venous line Neutropenia persistence Growth factors at death
0.019 0.024 <0.001a <0.001 <0.001 0.022 0.014 0.021
a
Significant on multivariate analysis.
950
M. Ellis / Molecular Immunology 38 (2001) 947–957
Fig. 3. CT halo sign.
3.2. The computerized tomographic (CT) halo sign This is now recognized as the radiologic hallmark of invasive pulmonary aspergillosis (IPA; Caillot et al., 1997), in the patient with hematological malignancy, since it has been validated by histopathologic sampling. Its angioinvasiveness is the pathological basis for the glow-blush of contrast around the central area (Fig. 3). The differential diagnosis of this radiologic halo sign includes mucormycosis, tumor cell infiltration, candidal lung infection, granulomatous lung diseases, tuberculoma and pulmonary hemorrhage. However, in this population with low risk for the other possibilities, the halo sign remains an extremely valuable non-invasive diagnostic tool, highly predictive for IPA. It has been suggested that detection of IA through serial computerized tomographic (CT) scanning is significantly associated with an improved survival for patients with IA (Caillot et al., 1997). Such an early diagnostic approach was linked to early administration of moderately high dose liposomal amphotericin B (LAB) treatment, at Tawam Hospital. Between 1998 and 2000, 180 patients with acute leukemia/myelodysplastic syndrome/lymphoma underwent induction/consolidation chemotherapy. Fifty-three of the 180 (29%) developed neutropenic fever and were at risk for IFI. CT and plain chest radiography were undertaken on the 4th febrile day and repeated weekly. The diagnostic criteria for probable IPA were: ARNF + <0.1 × 109 neutrophils/l + pulmonary signs + negative bacteriology + negative indications for alternative fever etiologies + positive CT scan evidence of IA. On this basis 21 of the 53 patients had probable IA. These patients were treated with a mean dose of liposomal amphotericin B of 4.4 mg/kg/day for a median duration of 19 (4–57) days. The clinical response rate was 81%, crude mortality 43% and attributable mortality 9.5%. These findings compare
favorably with current reports of historical outcomes of 54, 67 and 28%, respectively (Caillot et al., 2001; Kim et al., 2001). A linked approach of early and serial CT scanning with moderately high doses of liposomal amphotericin B in neutropenic patients at high risk for IA improves the therapeutic outcome. The additional finding that the plain chest radiographs were either normal or non-specific at the time of the positive CT scans reinforces the need to optimize radiologic investigation. 3.3. Subcellular components Precision diagnostics are currently focused on the detection of fungal cell wall constituents and their antibodies and fungal nucleic acid in host body fluids. The PCR is oversensitive because of colonization-contamination but the use of a competitive PCR which raises the threshold for detection more closely relates to increased fungal load, reducing less false positives. Using universal primers to target conserved RNA gene sequences in all fungi—a panfungal PCR—is followed by exploiting the highly variable internal transcribed spacer regions through nested PCR to determine precise species (Sandhu et al., 1995). Cell wall antigen detection, e.g. galactomannan in Aspergillus is also oversensitive because of cross-reacting epitopes. However, it can be found early, before symptoms and signs, and titers fall if treatment is successful. Galactomannan may appear before PCR. In patients who survive, galactomannan levels fall in stark contrast to patients who die who have unchanged high levels (Maertens et al., 2001). Sensitivities and specificities for Aspergillus diagnosis are among the highest when a combination of ELIZA and PCR is used (Bretagne et al., 1998). The galactomannan sandwich ELIZA has had its credibility validated against histology/culture based gold standard
M. Ellis / Molecular Immunology 38 (2001) 947–957
reference points as a highly sensitive (90%) and specific (98%) non-invasive diagnostic tool recently. A potential benefit of successful application of this diagnostic test would be to rationalize the use of broad spectrum antifungals. It might be possible to reduce their use by 75% compared to use classical clinical diagnostic criteria. Cost, toxicity, emergent resistance could all be favorably influenced by selectively using antifungal drugs, based on positive serologic findings. However, a negative result cannot exclude Fusarium spp., Alternaria spp. and other molds—and so one could not withhold fungal treatment on the basis of a negative ELIZA result alone. A number of other issues need clarification. For example, there is a high false positive rate in the galactomannan assays—around 15%—the reason for this remains unclear. Dietary galactomannan translocation across damaged mucosa, cross-reactivity with Penicillium spp. and other antigens and partial treatment responses are all possible explanations. The other areas for investigation include optimal sampling frequency, the significance of galactomannan levels at different stages of infection and its usefulness to monitor and guide therapy.
4. Treatment of IFI 4.1. Antifungals The priority to define and use an antifungal agent having a substantial margin of safety is underscored from the adverse
951
experience with conventional amphotericin B (CAB), the well tried gold standard of care. Systemic toxicity occurs in 80% of patients. If CAB related renal failure occurs, the mortality in cancer patients triples to 54%, length of hospital stay increases to 8 days and there is a cost increment increase of US$ 30,000 per patient to treat the renal failure (Bates et al., 2001). On the other hand, attempts to reduce the incidence and severity of both general systemic reactions and local target organ toxicity, particularly renal, have in general been disappointing. Saline loading for example may not be tolerated in patients with hematologic malignancy, many of whom have cardiovascular compromization or increased pulmonary capillary permeability. Steroids and opiates, to reduce chills and rigors are used in around 50% of patients (Gallis et al., 1990). Their administration exposes the patient from further drug adverse reactions, which include heightening the immunocompromization status. Simple measures such as slowing the infusion rate may be beneficial in reducing severe chills and rigors. In a prospective double blind study of CAB given either over 1 or 4 h, the incidence of severe chills and rigors and tachycardia were significantly less with the slower infusion (Ellis et al., 1992; Fig. 4). Recently this has been extended to a 24 h continuous infusion (Eriksson et al., 2001). One of the postulated mechanisms for renal toxicity is CAB induced glomerular toxicity. Low dose dopamine has been investigated for its potential to prevent such a side effect, utilizing its physiologic property of increasing glomerular perfusion.
Fig. 4. Mean maximum pulse rise in two infusion rates of amphotericin B.
952
M. Ellis / Molecular Immunology 38 (2001) 947–957
Disappointingly a controlled study not only failed to show a significant protective effect but there was a high incidence of dopamine related adverse reactions (Camp et al., 1998). One of the most notable advances has been the reclothing of this emperor of antifungal therapy. Lipidisation had as its goal, an increase in the therapeutic index as a result of reduction in the toxic denominator and targeting of the AF at the site of fungal sepsis. Clearly this is an important strategic concept as there is considerable opportunity to correct the subtherapeutics arising from underdosing, dose interruption/discontinuation and inappropriate steroid, opiate and dopamine use. There are three commercially available lipid formulations of amphotericin B: ABLC, ABCD and LAB. These three differ considerably in structure, pharmacokinetics and pharmacodynamics. These aspects have been reviewed in detail elsewhere (Arikan and Rex, 2001). Each of the three compounds is unique in its properties; this is reflected in the difference in toxicity between them. The liposomal preparation (LAB) currently appears to have documented superior performance in this respect. A randomized study by Wingard et al. (2000) showed a significant less incidence in chills, rigors, nephrotoxicity and therapy discontinuations with LAB compared to ABLC. The efficacy impact of these highly expensive compounds has been the subject of intense debate (Ellis, 2000). Dose for dose, animal models demonstrate they have better mycologic eradication and survival compared to CAB. For example, there is less hematogenous dissemination from primary lung foci, increased lung sterilization and improved survival when the lipid compounds are compared to CAB (Leenders et al., 1996; Allende et al., 1994). A vast amount of information is also provided by open label human studies which include salvage treatment situations. These confirm high complete response rates with cure rates of generally around 60–70% for IA when various lipid formulations are used (Ellis, 2000). Controlled human clinical trials confirm the heightened efficacy of LAB over CAB. For example, there is improved fever resolution in patients with ARNF (Prentice et al., 1997), reduction in breakthrough IFI in patients at risk (Walsh et al., 1999) and a better response in the treatment of patients with established IFI (Leenders et al., 1998). Despite over 40 years of experience with amphotericin B there is no clear evidence which defines an optimal dose. Careful scrutiny of results from animal models reveals an interesting phenomenon. Mycologic eradication is clearly better at higher dosages, but survivorship does not parallel this to anything like the same degree— “mycologic-survival disassociation” (Ellis, 2000). This is clearly seen in Allende’s study (1994) where survivorship in the granulocytopenic rabbit infected with Aspergillus is 60% for 1 and 5 mg/kg/day of ABCD and 25% for 10 mg. Further support for this has been forthcoming in the only randomized controlled clinical trial of antifungal dosing
ever to be performed. EORTC #19923 showed that 1 and 4 mg of LAB in treating probable and proven IA were equi-effective in response and survival (Ellis et al., 1998). This apparently surprising finding has prompted many questions. The fact that a small group of patients with biopsy proven IA tended to do better with the higher dose suggested the trial may not have been powered sufficiently to show any significant difference at the higher dose. Perhaps a much higher dose differential was needed to be effective. Cenciu’s (1997) animal model in which a maximum tolerated dose was 100% successful suggests that might be the case. The ability of the controlled clinical trial as a model for investigating antifungal efficacy has recently been questioned (Rex et al., 2001). Patient accrual is generally poor for reasons which include poor patient performance status, reluctance of physicians to depart from time-honored antifungal treatments and an increasing number of new antifungals which seriously dilute trial patient base numbers. There are also many confounding variables in these very ill patients, which could influence outcome measures and which are difficult to control. Finally there is a widespread high degree of diagnostic variability—one review of over 170 studies gave a low κ score in this regard (Ascioglu et al., 2001). Carefully performed in depth case studies will be helpful in solving some problems such as dosing issues. For example, Patterson (2000) showed that in a patient that died from IPA, the antifungal drug concentrations were all low and below the MICs in the infected lung tissue but above the MICs in the adjacent healthy tissue. This suggested that poor outcome in patients with invasive mycoses may be in part related to failure to effectively deliver antifungals to the site of infection. 4.2. Recent antimycotic drugs A plethora of AF drugs with cell wall and subcellular targets other than the cell membrane have appeared in recent years. There has been a radical change in philosophy from “find an antifungal and then find a fungus to treat” to “investigate the genome for virulence factors and synthesize an antifungal to target those”. Among the third generation azoles is voriconazole whose target is the 14-␣-demethylase enzyme responsible for ergosterol synthesis of the cell wall, with remarkable broad spectrum activity to yeasts and molds. Posaconazole is another new generation azole with substantial (75% response in salvage therapy) effect against one of the most pan-resistant fungi—Fusarium spp. The candins which inhibit 1,3--glucan are remarkable for lack of any cross-resistance with other antifungals. Among these is capsofungin, recently licensed for the treatment of refractory IA. Others include the pradimicins which bind to mannosides and disrupt fungal cell wall activity, nikkomycin chitin inhibitors and the sordarins which inhibit fungal elongation factor 2.
M. Ellis / Molecular Immunology 38 (2001) 947–957
4.3. Exploitation of existing antifungals Molecular-stearic remodeling has already been exemplified with the lipid formulations of CAB. Nystatin has also been reformulated as a liposomal preparation, with major toxicity reduction. Liposomal forms of miconazole and ketoconazole are also under investigation. Recently immunoliposomes have been created. These contain fungus specific antibodies to facilitate directed fungal cell targeting. Complexing lipsosomal amphotericin B with polyethylene glycol (PEG-LAB) substantially increases the circulating half life and target fungi for longer periods of time. However, these preparations appear to yield less intracellular concentration of amphotericin B compared with the parent CAB, which may translate into inferior efficacy in the long term. Cochleates, lipid based supramolecular assemblies containing negatively charged phospholipid, natural products and divalent cations are among the latest innovative formulation vehicles to undergo clinical testing. Combination therapy of two different antifungals in general has not produced clear clinical advantages, apart from either fluconazole or amphotericin B in combination with 5-fluorocytosine which have synergistic activity against Cryptococcus neoformans meningitis. Combination of a primary antifungal with an agent having a different action has produced some important observations. For example, an important mechanism in Candida resistance is efflux pumping of azoles out of the yeast cell prior to target binding. White (1997) demonstrated in HIV patients that as the duration of the use of fluconazole increased over several months, mRNA for the efflux pump was progressively expressed which paralleled the rise in MICs and use of higher fluconazole doses. The use of MC-510027 an efflux pump inhibitor has now been shown to reduce the MICs of the azoles in such situations (Chamberland et al., 1999).
5. Experimental strategies It is crucial that therapeutic modalities are explored which have interventional characteristics other than direct antimicrobial. Fungal resistance, clinical response failures in the presence of fully sensitive organisms, overwhelming acute onset septic states that occur before an antifungal can be reasonably expected to work—are some of the compelling reasons. Current experimental approaches include the following. 5.1. Hyperbaric oxygen (HBO) This unusual approach harvests the demonstrable properties of oxygen under increased tension such as augmentation of white cell mediated fungal killing, optimization of tissue pO2 and reduction of edema and acidosis. Substantial improvements in outcome have been documented for
953
selected fungal infections such as mucormycosis (Price and Stevens, 1980). Fig. 5 shows a patient with hematogenous IA whose large necrotic skin lesions did not respond to LAB until hyperbaric oxygen (HBO) was added as adjunctive therapy. 5.2. Surgical adjunctive therapy The surgical management of IFI aims to substantially reduce large fungal masses or to remove dormant lesions which could re-activate during future immunosuppression. In patients with hematologic malignancy and IPA, resection of well circumscribed lung masses during the post neutrophil recovery phase reduces future chemotherapy induced reactivation of the IPA by 50% (Karp et al., 1988). Mortality from catastrophic bleeding associated with the mycotic lung sequestrum of lung cavitary aspergillosis can also be reduced by timely surgical intervention in these patients (Wong et al., 1992). Fungal endocarditis is another illustration of the mandatory use of adjunctive surgical intervention. A meta-analysis of the world literature indicated that outcome in such patients was improved when antifungal treatment was combined with surgical valve replacement (Ellis et al., 2001). This approach appears to be the most beneficial for patients with Aspergillus endocarditis as mortality is virtually 100% with amphotericin B treatment alone. 5.3. Targeting virulence-pathogenic factors The phospholipases of C. albicans were first recognized 40 years ago. These are produced in great quantities shed from the very tips of the hyphae and are correlated with tissue invasion and cell cytolyis. Parallelism exists between phospholipase production, hematogenous dissemination and mortality (Ghannoum, 2000). At least three genes have been identified. Not only do the phospholipases permit host cell entry, but they deregulate cell signaling, induce IL-8/-6 expression and induce inflammatory mediators such as aracidonic acid. Synthetic phospholipase inhibitors have been manufactured which prolong animal survival when used with fluconazole to treat severe sepsis (Hanel et al., 1995). 5.4. Neutrophil properties One of the pivotal requirements for recovery from IFI is correction of neutropenia. The normal observation in a patient with IA and neutropenia is that recovery of neutropenia is mandatory for effective recovery from the IA, despite antifungal agents. Previous concerns that white cell transfusions posed major problems of safety and efficacy have been largely quelled with current state of the art cell harvest technology, donor screening for blood borne infections and assurance of good quality and quantity of white cells made possible by the use of recombinant colony stimulating factors. There is resurrected interest in granulocyte transfusions to support such patients with severe IFI (Price et al., 2000).
954
M. Ellis / Molecular Immunology 38 (2001) 947–957
Fig. 5. (a) Cutaneous hematogenous aspergillosis before HBO treatment: note ischemia and edema; (b) same patient after HBO treatment: note resolution of signs and healthy granulation tissue.
It is the primary azurophil granules and the secondary granules of the neutrophil that contain a variety of naturally occurring antimicrobial peptides and proteins and contribute towards the host’s innate immune response. These are deployed at the sites of fungal infection. Seroprocidin in one example—it is in fact a family which produce lethal membrane perturbations (Levy, 2000). The possibility of utilizing recombinant preparations is now a reality facing clinical testing. 5.5. Soluble immune system enhancement This appears to be of crucial importance in the successful response to IFI in the immunocompromized. Unfortunately, the rapidity with which colony stimulating factors were introduced into clinical practice and are now widely
used almost on a routine basis precludes a pure evidence based evaluation of their efficacy. There is good circumstantial evidence however for an improved outcome when granulocyte, monocyte growth factors, interferon-␥, IL-12, TNF-␣ and others are used as adjunctive therapy. The ECOG study of Rowe (1998) showed significant improved survivorship when colony stimulating factors were used in addition to antifungal treatment of IFI. Immunomanipulation may also be beneficial in the “not so immunocompromized patient” with an invasive mycosis. An example of this phenomenon is invasive sinus aspergillosis usually due to Aspergillus flavus, characterized by progressive erosive and destructive skull base disease and intracranial spread from an ethmoid sinus epicenter (deShazo et al., 1997). Many such patients have no gross or apparent immunodeficiency. One such patient had subtle immune
M. Ellis / Molecular Immunology 38 (2001) 947–957
955
Fig. 6. (a) Extensive intracranial aspergillosis, spreading through cribriform plate (MRI SE T1 weighted image); (b) same patient after combination antifungal and immunomodulatort treatment showing extensive disease reduction.
defects noted (Ellis et al., 2002). These were: CD4-penia (458 cells/l), low percentages of CD3+ , CD15+ , and natural killer cells (5%), weak quantitative nitroblue tetrazolium reduction responses (31–50% normal) and high IgE levels (433 IU/ml) in the absence of allergy. Taken together these findings suggested a selective Th1 cell defect, which could have predisposed the patient to invasive aspergillosis. She presented with a massive intracranial A. flavus infection which included MRI-scan documentation of infiltration of the optic nerves and visual loss. Despite up to 10 mg/kg/day of liposomal amphotericin B she deteriorated. She did not respond until GMCSF and ␥-interferon were introduced as part of the therapy, consequently clearing most of the intracerebral fungal load (Fig. 6). GMCSF is known to enhance neutrophil, monocyte and macrophage antifungal activities in vivo, and ␥-interferon improves antifungal properties of polymorphonuclear cells and of helper T cells by directing their activities towards a Th1 response and inducing neutrophil mediated fungal hyphal damage. The case study underscores on the one hand our limited understanding of the precise fungal–host interactive mechanisms in such patients and on the other, despite this, such patients appear to respond to combination antifungal-immunomodulatory treatments.
Further clarification in this field will define subpopulations more likely to benefit from such strategies. 5.6. Prophylactic cytokines Finally, a recent interesting concept is that of the cytokine shield. The majority of invasive yeasts and bacteria arise from the gut, translocating as a result of chemotherapy induced apoptosis of the gut epithelium with consequent pathologica changes in gut wall permeability. Recombinant IL-11 used for managing thrombocytopenia also has epithelial growth potential. The hypothesis is that if gut integrity can be preserved through the chemotherapy treatment phase, fungal and bacterial infections might be reduced in frequency and severity. A neutropenic mouse model showed that survival following a lethal dose of Pseudomonas aeruginosa could be increased with r IL-11 (Opal et al., 1999). A double blind placebo controlled trial of recombinant IL-11 given to leukemic patients to assess its effect on preventing sepsis is underway. An interim safety analysis of the first 20 patients, whilst preserving anonymity of the trial drugs has indicated a difference in some of the end points of the two study arms. Septic events, fungal events, severity of sepsis
956
M. Ellis / Molecular Immunology 38 (2001) 947–957
Table 2 Recombinant interleukin-11 in leukemic patients Preliminary crude data on 20 patients
Patients (n) Total infection events (n) Sepsis score ≥ 3a Any bacteremiaa Cryptic fungal infectionsa Abnormal lactulose: mannitol ratioa Lactulose:mannitol ratio (median) Adverse eventsa
Drug A
Drug B
11 11 3 8 6 4b
9 7 1 4 2 2c
0.12
0.02
3 (injection pain, injection pain, edema)
2 (edema, pleural effusion)
Normal lactulose:mannitol ratio < 0.01. a Number of patients. b Associated with bacteremias due to Pseudomonas, Enterococcus, Staphylococcus, Klebsiella, and Fusobacterium spp. c Associated with bacteremia due to Klebsiella spp.
and intestinal epithelial cell integrity (as assessed with urinary lactulose:mannitol ratios) were all consistently different between the two arms (Table 2). Should the ultimate findings suggest an advantage for recombinant IL-11 in reducing fungal and bacterial sepsis, this would give further encouragement for exploration of such non-antimicrobial based strategies in managing invasive mycoses. References Abi-Said, D., Anaissie, E., Uzun, O., et al., 1997. The epidemiology of hematogenous candidiasis caused by different Candida species. CID 24, 1122–1128. Allende, M.C., Lee, J.W., Francis, P., et al., 1994. Dose dependent antifungal activity and nephrotoxicity of amphotericin B colloidal dispersion in experimental pulmonary aspergillosis. Antimicrob. Agents Chemother. 38, 518–522. Arikan, S., Rex, J., 2001. Lipid-based antifungal agents: current status. Curr. Pharma. Des. 7, 393–415. Ascioglu, S., de Pauw, B., Donnely, J., et al., 2001. Reliability of clinical research on invasive fungal infections: a systematic review of the literature. Med. Mycol. 39, 35–40. Bates, D., Su, L., Yu, D., et al., 2001. Mortality and costs of acute renal failure associated with amphotericin B therapy. CID 32, 686–693. Bretagne, S., Costa, J., Bart-Delabesse, E., et al., 1998. Comparison of serum galactomannom antigen detection and competitive polymerase chain reaction for diagnosing invasive aspergillosis. CID 26, 1407– 1412. Caillot, D., Casasnovas, O., Bernard, A., et al., 1997. Improved management of invasive pulmonary aspergillosis in neutropenic patients using early thoracic computed tomographic scan and surgery. J. Clin. Oncol. 15, 139–147. Caillot, D., Couaillier, J., Bernard, A., et al., 2001. Increasing volume and changing characteristics of invasive pulmonary aspergillosis on sequential thoracic computer tomography scans in patients with neutropenia. J. Clin. Oncol. 19, 253–259. Camp, M., Wingard, J., Gilmore, C., et al., 1998. Efficacy of lowdose dopamine in preventing amphotericin B nephrotoxicity in bone marrow transplant patients and leukemia patients. Antimicrob. Agents Chemother. 42, 3103–3106.
Cenciu, E., Perito, S., Enssle, K., et al., 1997. Th1 and Th2 cytokines in mice with invasive aspergillosis. Infect. Immunol. 65, 564–570. Chamberland, S., Blais, J., Cotter, D., 1999. Impact of MC-510,027: a fungal efflux pump inhibitor, on the susceptibility of clinical isolates of Candida spp. to antifungal agents. In: Proceedings of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, September 1999. Abstract #1270. deShazo, R., O’Brier, M., Chapin, K., et al., 1997. A new classification and diagnostic criteria for invasive fungae sinusitis. Arch. Otolarygol. Head Neck Sing. 123, 1181–1188. Ellis, M., 2000. Amphotericin B preparations: a maximum tolerated does in severe invasive fungal infections? Transplant Infect. Dis. 2, 51–61. Ellis, M., al-Hokail, A., Clink, H., et al., 1992. Double blind randomized study of the effect of infusion rates on toxicity of amphotericin B. Antimicrob. Agents Chemother. 36, 172–179. Ellis, M., Qadri, S., Spence, D., et al., 1994. The effect of fluconazole as prophylaxis for neutropenic patients on the isolation of Candida species from surveillance cultures. J. Antimicrob. Chemother. 33, 1223–1228. Ellis, M., Spence, D., Bouchama, A., et al., 1995. Open lung biopsy provides a higher and more specific diagnostic yield compared to broncho-alveolar lavage in immunocompromised patients. Scand. J. Infect. Dis. 27, 157–162. Ellis, M., Spence, D., de Pauw, B., et al., 1998. An EORTC international multicenter randomized trial (EORTC number 19923) comparing two dosages of liposomal amphotericin B for treatment of invasive aspergillosis. CID 27, 1406–1412. Ellis, M., Al-Abdely, H., Sandridge, A., et al., 2001. Fungal endocarditis: evidence in the world literature 1965–1995. CID 32, 50–62. Ellis, M., Watson, R., McNabb, A., et al.., 2002. Massive intracerebral aspergillosis responding to combination high dose liposomal amphoetricin B and cytokine therapy without surgery. J. Med. Microbiol. 51, 70–75. EORTC, International Antimicrobial Therapy Cooperative Group, 1989. Empirical antifungal therapy in febrile granulocytopenic patients. Am. J. Med. 86, 668–672. Eriksson, U., Seifert, B., Scaffner, A., 2001. Comparison of effects of amphotericin B deoxycholate infused over 4 or 24 h: randomized controlled trial. BMJ 322, 579–582. Gallis, H., Drew, R., Pickard, W., 1990. Amphotericin B: 30 years of clinical experience. Rev. Infect. Dis. 12, 308–329. Gerson, S., Talbot, G., Hurwitz, S., et al., 1985. Discriminant scorecard for diagnosis of invasive pulmonary aspergillosis in patients with acute leukemia. Am. J. Med. 79, 57–64. Ghannoum, M., 2000. Potential role of phospholipases in virulence and fungal pathogenesis. Clin. Microbiol. Rev. 13, 122–143. Groll, A., Shah, P., Mentzel, C., et al., 1996. Trends in the postmortem epidemiology of invasive fungal infections at a university hospital. J. Infect. 33, 23–32. Gunsilius, E., Lass-Florl, C., Kahler, C., 2001. Candida ciferrei, a new fluconazole-resistant yeast causing systemic mycosis in immunocompromized patients. Ann. Hematol. 80, 178–179. Hanel, H., Kirsch, R., Schmidts, H., et al., 1995. New systematically active antimycotics from the -blocker category. Mycoses 38, 251–264. Jarvis, W., 1995. Epidemiology of nosocomial fungal infections, with emphasis on Candida species. CID 20, 1526–1530. Karp, J., Burch, P., Merz, W., 1988. An approach to intensive antileukemia therapy in patients with previous aspergillosis. Am. J. Med. 85, 203– 206. Kim, M., Lee, K., Kim, J., et al., 2001. Crescent sign in invasive pulmonary aspergillosis: frequency and related CT and clinical factors. J. Comput. Assist. Tomogr. 25, 305–310. Levy, O., 2000. Antimicrobial proteins and peptides of blood: templates for novel antimicrobial agents. Blood 96, 2664–2672. Leenders, A.C.A.P., De Marie, S., Ten Kate, M.T., Bakker-Woudenberg, I.A.J.M., Verbrugh, H.A., 1996. Liposomal amphotericin B (Ambisome) reduces dissemination of infection as compared with amphotericin B deoxycholate (Fungizone) in a rat model of pulmonary aspergillosis. J. Antimicrob. Chemother. 38, 215–225.
M. Ellis / Molecular Immunology 38 (2001) 947–957 Leenders, A.C.A.P., Daenen, S., Jansen, R.L.H., et al., 1998. Liposomal amphotericin B compared with amphotericin B deoxycholate in the treatment of documented and suspected neutropenia associated invasive fungal infections. Br. J. Haematol. 103, 205–212. Maertens, J., Verhaegen, J., Lagrou, K., et al., 2001. Screening for circulating galactomannan as a noninvasive diagnostic tool for invasive aspergillosis in prolonged neutropenic patients and stem cell transplantation recipients: a prospective evaluation. Blood 97, 1604– 1610. Opal, S.M., Jhung, J.W., Keith, J.C., et al., 1999. Additive effects of human recombinantinterleukin-11 and granulocyte colony-stimulating factor in experimental gram-negative sepsis. Blood 93, 3467–3472. Patterson, P., Johnson, B., Ainscough, S., et al., 2000. Treatment failure in invasive aspergillosis due to poor tissue penetration by antifungal agents. In: Proceedings of 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, September 2000. Abstract #1328. Pfaller, M., Jones, R., Doern, G., et al., 1998a. International surveillance of bloodstream infections due to Candida species: frequency of occurrence and antifungal susceptibilities of isolates collected in 1997 in the United States, Canada and South America for the SENTRY programme. The SENTRY participant group. J. Clin. Microbiol. 36, 1886– 1889. Pfaller, M., Jones, R., Messer, S., et al., 1998b. National surveillance of nosocomial blood stream infection due to Candida albicans: frequency of occurrence and antifungal susceptibility in the SCOPE program. Diagn. Microbiol. Infect. Dis. 31, 327–332. Powderley, W., Finkelstein, D., Feinberg, J., et al., 1995. A randomized trial comparing fluconazole with clotrimazole troches for the prevention of fungal infections in patients with advanced human immunodeficiency virus infection. NEJM 332, 700–705. Prentice, H.G., Hann, I.M., Herbrecht, R., et al., 1997. A randomised comparison of liposomal versus conventional amphotericin B for the treatment of pyrexia of unknown origin in neutropenic patients. Br. J. Haematol. 98, 711–718. Price, J., Stevens, D., 1980. Hyperbaric oxygen in the treatment of rhinocerebral mucormycosis. Laryngoscope 90, 737–747.
957
Price, T., Bowden, R., Boeckh, M., et al., 2000. Phase I/II trial of neutrophil transfusions from donors stimulated with G-CSF and dexamethasone for treatment of patients with infections in hematopoietic stem cell transplantation. Blood 95, 3302–3310. Rees, J.R., Pinner, R., Hajjeh, R., et al., 1998. The epidemiological features of invasive mycotic infections in the San Francisco Bay area, 1992–1993: results of population-based laboratory active surveillance. CID 27, 1138–1147. Rex, J., Walsh, T., Nettleman, M., et al., 2001. Need for alternative trial designs and evaluation strategies for therapeutic studies on invasive mycoses. CID 33, 95–106. Rowe, J., 1998. Treatment of acute myeloid leukemia with cytokines: effect on duration of neutropenia and response to infections. CID 26, 1290–1294. Sandhu, G., Kline, B., Stockman, L., et al., 1995. Molecular probes for diagnosis of fungal infections. J. Clin. Microbiol. 33, 2913–2919. Verweij, P., Donnelly, J., de Pauw, B., et al., 1996. Prospects for the early diagnosis of invasive aspergillosis in the immunocompromised patient. Rev. Med. Microbiol. l7, 105–113. Viscoli, C., Girmenia, C., Marinus, A., et al., 1999. Candidemia in cancer patients: a prospective, multicenter surveillance study by the Invasive Fungal Infection Group (IFIG) of the European Organisation for Research and Treatment of Cancer (EORTC). CID 28, 1071–1079. Walsh, T.J., Finberg, R.W., Arndt, C., et al., 1999. Liposomal amphotericin B for empirical therapy in patients with persistent fever and neutropenia. N. Engl. J. Med. 340, 764–771. White, T., 1997. Increased mRNA levels of ERG16, CDR, and MDR1correlate with increases in azole resistance in Candida albicans isolates from a patient infected with human immunodeficiency virus. Antimicrob. Agents Chemother. 41, 1482–1487. Wingard, J., White, M., Anaissie, E., et al., 2000. A randomized double-blind comparative trial evaluating the safety of liposomal amphotericin B versus amphotericin B lipid complex in the empirical treatment of febrile neutropenia. CID 31, 1155–1163. Wong, K., Waters, C., Walesby, R., 1992. Surgical management of invasive pulmonary aspergillosis in immunocompromised patients. Eur. J. Cardio-Thorac. Surg. 6, 138–143.
Invasive fungal infections: evolving challenges for diagnosis and therapeutics Michael Ellis∗ Faculty of Medicine and Health Sciences and Tawam Hospital, Al Ain, Abu Dhabi, United Arab Emirates Received 15 October 2001; accepted 10 December 2001
Abstract Invasive fungal infections (IFI) parallel the explosive increase in the immunocompromized patient population, and are characterized by diagnostic difficulties and extreme mortality. Candidemia in a tertiary referral hospital in the Middle East confirms the current epidemiologic shift in this common blood stream pathogen towards non-malignancy cases (38%) and antifungal prophylaxis failure (20%), high presentation sepsis scores and attributable mortality (32%). Invasive aspergillosis (IA) is also associated with high mortality. Use of non-invasive computerized tomographic (CT) radiologic scanning linked to early administration of high dose liposomal amphotericin B (LAB) is associated with a reduced mortality of 9.5% compared to historical experience of 28%. Life threatening invasive aspergillosis also occurs in patients who are less obviously immunocompromized. Investigations may reveal subtle immune deficits which could place the patient at some risk for an invasive mycosis. Antifungal treatment used in combination with progenitor cell growth factors and ␥-interferon has proved successful in such situations of progressive fungal disease unresponsive to antifungal therapy alone. Pharmacologic remodeling of existing compounds by lipidisation reduces both the toxicity denominator and the efficacy numerator of the therapeutic index when compared to the parent drug. A comparative dose study of liposomal amphotericin B in aspergillosis has demonstrated equi-efficacy, generated debate over the ability of the controlled clinical trial to be capable of assessing antifungal efficacy, and illustrated that recovery from an invasive fungal infection may require maximum tolerated doses and immunomanipulation. Several new antifungal strategies are under clinical investigation. These include reformulating existing antifungals, exploitation of the growing knowledge of virulence factors to synthesize antagonists, immune reconstitution and immunoprotection. An interim analysis of an ongoing placebo controlled study of recombinant interleukin-11 to assess its efficacy in reducing sepsis in leukemia patients through prevention of chemotherapy induced gut epithelial cell apoptosis, has demonstrated a difference in the two study arms in sepsis rates and preservation of gastrointestinal epithelial cell integrity. The unique and special challenges presented by the dynamic epidemiologics of invasive fungal infections are demanding and attracting considerable responses, in the fields of diagnosis and therapeutics. Current strategies need considerable improvement, yet ongoing collaborative efforts will have a positive impact on our understanding of the fungus–host interaction and ultimately our ability to offer better care for our patients with invasive mycoses. © 2002 Published by Elsevier Science Ltd. Keywords: Candida; Aspergillus; Amphotericin B; Immunocompromization; Halo sign; Cytokines
1. Introduction
1.1. The scope of the problem
This article focuses on the current status of invasive fungal infections (IFI) from a clinical perspective. It reviews the major challenges in diagnostics and treatments and includes local case studies and preliminary results of clinical studies conducted at Tawam Hospital, UAE.
The end of the 20th century witnessed a remarkable change in the pattern of pathogens responsible for serious human infections. Among hospitalized patients fungal etiology of pneumonia, urinary tract infection, surgical wound infection and blood stream infections had increased 50, 100, 200 and 400%, respectively from 1990 through 2000 (Jarvis, 1995). Candidemia had risen to rank 4th among blood stream pathogens in the US (Abi-Said et al., 1997), its
∗ Fax: +971-3-722-3942. E-mail address: [email protected] (M. Ellis).
0161-5890/02/$ – see front matter © 2002 Published by Elsevier Science Ltd. PII: S 0 1 6 1 - 5 8 9 0 ( 0 2 ) 0 0 0 2 2 - 6
948
M. Ellis / Molecular Immunology 38 (2001) 947–957
lethality reflected in the high crude mortality rate of 40%. A unique population based surveillance study in 1998 from San Francisco gave good evidence for the first time that IFI were not just a phenomenon confined to more severely immunocompromized hospital patients, but that the problem was endemic throughout the general population (Rees et al., 1998). One of the surprising findings from that study was the exponential explosive rise of invasive candidiasis with age—an indicator also that IFI do not occur solely in the more seriously or classically immunocompromized.
2. Epidemiology and causality A heterogeneity of factors is responsible for the emergence of invasive mycoses as a significant cause of morbidity and mortality. These include: the increased bipolarity of age at the one end characterized by a striking increase in pre-term births (16% increase in 10 years in some US population) and at the other by the significant increased pool of the “super-old”, socio-behavioral disintegration associated with HIV, increased survivorship of people with congenital immunodeficiency syndromes, rigorous and repetitive abdominal surgery, widespread use and abuse of antimicrobials, intrusive vascular lines, transmucosal devices and total parental nutrition support systems. It is simply not the disproportionate increase in IFI that is occurring but also a notable epidemiologic shift in the types of molds and yeasts. The relative contribution of Aspergillus, as shown in autopsy studies has risen steadily, replacing Candida as the top pathogen in some institutions (Groll et al., 1996). Fusarium was once confined to an uninteresting dermatophyte role—now it has moved out of its ecologically secure niche to become second to Aspergillus as a major invasive mold, mimicking aspergillosis clinically and radiologically. Penicillium spp. are in general innocuous, but now life threatening TB-like infections are described; penicilliosis has became the 3rd commonest AIDS defining infection in Thailand. In 1963 only five Candida spp. were known to cause human disease—now more than 20 yeasts are commonly reported including Malassezia, Rhodotorula and Prototheca spp. Among the invasive Candida spp. the non-albicans species have emerged, accounting for up to 60% of all the candidemias seen in some hospitals (Pfaller et al., 1998a,b)—the most recent being Candida ciferrei with marked azole resistance (Gunsilius et al., 2001)—bringing challenges for effective antifungal therapy. This epidemiologic shift is not clearly understood. The widespread use of azoles has been implicated (Viscoli et al., 1999) but a causal relation remains unproven. Retrospective and uncontrolled data analysis, single center sources (institution bias), nosocomial outbreaks (substantiated by molecular epidemiologics), increase in non-albicans which pre-date azole use and greatly increased chemotherapeutic regimens
Fig. 1. Proportion of C. albicans: non-albicans isolated from surveillance swabs during a 3-week period of fluconazole prophylaxis.
are important confounding variables which could also contribute to this shift. Furthermore the epidemiologic data of C. krusei (an intrinsically azole resistant species) as a surrogate marker for the azole drive to emergence of resistant non-albicans species, does not support this supposition. Thus there is a constant, rather than a rising, incidence of this species over many years at oncology centers and other large populations (e.g. HIV infected persons) in which azole use is high (Powderley et al., 1995). Nevertheless some prospective analyses do show that superficial fungal colonization (in contrast to invasive candidiasis) is subject to azole pressure. During 3 weeks of fluconazole prophylaxis, the proportion of non-albicans Candida spp. isolated from throat, urine and stool surveillance cultures substantially and significantly increased (Ellis et al., 1994; Fig. 1). Yet the invasive non-albicans Candida isolates occurred only in those patients not receiving fluconazole prophylaxis. These observations suggest other factors additional (or in spite of) azole use influence the current shift towards non-albicans candidiasis. The local experience at Tawam Hospital reflects the hospital immunocompromized scene with regard to candidiasis. Over a 5-year period there have been 60 episodes of candidemia. The expected association with hematologic malignancy is seen but the shift towards non-malignant conditions such as repetitive abdominal surgery and prematurity is striking (Fig. 2a). The other well recognized risk factors are confirmed but the high association with recent bacteremias is a novel finding. Breakthrough candidemia which should have been prevented by effective antifungal prophylaxis occurred in 20% of the patients and is a cause for concern. Over half of the isolates were non-albicans spp. (Fig. 2b). High sepsis scores and multi-organ failure dominated the clinical presentation (Table 1A). The mortality rate of 50%
M. Ellis / Molecular Immunology 38 (2001) 947–957
949
3. Current diagnostic tools Classic blood culture proves positive in less than 5% of patients in invasive aspergillosis (IA) and in 50% in invasive candidiasis. Invasive techniques such as open lung biopsy currently provide the greatest and most accurate diagnostic yield. In a unique prospective study in which two invasive diagnostic techniques were concurrently performed the yield of specific diagnosis made by open lung biopsy was three-fold greater than bronchoalveolar lavage (Ellis et al., 1995). It is important to note that the range of diagnoses is wide and infections contribute only a proportion of the etiologies in neutropenic fever patients. However, such invasive techniques to prove an invasive mycosis at the point of target organ impact is very often impossible to undertake because of the critical coagulopathic clinical profile of these patients. Currently, the diagnostic approach centers on the following three areas. 3.1. High diagnostic acumen
Fig. 2. Candidemia at Tawam Hospital: (a) underlying disease categories; (b) Candida species isolated.
gives credence to the current clinical practice guidelines that ALL candidemias should be energetically managed. Of all variables tested, shock, organ failure, persistent neutropenia and use of growth factors were independent factors significantly predictive of death (Table 1B). The patient with leukemia and lymphoma is the most immunologically disprivileged since the neutrophil which forms the pivotal host defence has been removed and its function impaired through disease and chemotherapy. The fact that 25% of patients with hematologic malignancy die with IFI—many undiagnosed in life—underscores our inability to diagnose and manage effectively.
In a profound protracted neutropenic febrile patient, with no bacterial sepsis or any other cause for fever and no clinical focus, who is failing to respond to broad spectrum antibiotics (antibiotic resistant neutropenic fever or ARNF), IA or other IFI has to be strongly suspected (EORTC, 1989). Currently serologic diagnostics are increasingly available, but these have to be validated thoroughly in the clinical setting. In addition they are expensive to use for daily testing. The bedside clinical science therefore remains of paramount importance. This was underscored by Gerson et al. (1985) who combined 11 clinical parameters into a discriminant scorecard to distinguish patients with IA from controls. Retrospective and prospective phased trials gave sensitivities of 63–93% and specificities of 88–98%. These figures are virtually identical to those currently reported for Aspergillus galactomannan testing (Verweij et al., 1996). The diagnosis of IA using the clinical scorecard was made 4 days earlier compare to a more classical clinical approach. Patients with the untreated ARNF syndrome are at a high risk for rapid overt IFI.
Table 1 Candidemia (A) Characteristics at presentation (Tawam Hospital)
(B) Variables associated with death
Characteristics
Percentage
Variable
P
Shock Renal failure Multi-organ failure Sepsis score ≥ 3 Mortality crude Attributable Associated
43 23 33 68 50 32 13
ICU admission Malignant disease non-remission Shock at presentation Renal failure at presentation Multi-organ failure at presentation Non-removal central venous line Neutropenia persistence Growth factors at death
0.019 0.024 <0.001a <0.001 <0.001 0.022 0.014 0.021
a
Significant on multivariate analysis.
950
M. Ellis / Molecular Immunology 38 (2001) 947–957
Fig. 3. CT halo sign.
3.2. The computerized tomographic (CT) halo sign This is now recognized as the radiologic hallmark of invasive pulmonary aspergillosis (IPA; Caillot et al., 1997), in the patient with hematological malignancy, since it has been validated by histopathologic sampling. Its angioinvasiveness is the pathological basis for the glow-blush of contrast around the central area (Fig. 3). The differential diagnosis of this radiologic halo sign includes mucormycosis, tumor cell infiltration, candidal lung infection, granulomatous lung diseases, tuberculoma and pulmonary hemorrhage. However, in this population with low risk for the other possibilities, the halo sign remains an extremely valuable non-invasive diagnostic tool, highly predictive for IPA. It has been suggested that detection of IA through serial computerized tomographic (CT) scanning is significantly associated with an improved survival for patients with IA (Caillot et al., 1997). Such an early diagnostic approach was linked to early administration of moderately high dose liposomal amphotericin B (LAB) treatment, at Tawam Hospital. Between 1998 and 2000, 180 patients with acute leukemia/myelodysplastic syndrome/lymphoma underwent induction/consolidation chemotherapy. Fifty-three of the 180 (29%) developed neutropenic fever and were at risk for IFI. CT and plain chest radiography were undertaken on the 4th febrile day and repeated weekly. The diagnostic criteria for probable IPA were: ARNF + <0.1 × 109 neutrophils/l + pulmonary signs + negative bacteriology + negative indications for alternative fever etiologies + positive CT scan evidence of IA. On this basis 21 of the 53 patients had probable IA. These patients were treated with a mean dose of liposomal amphotericin B of 4.4 mg/kg/day for a median duration of 19 (4–57) days. The clinical response rate was 81%, crude mortality 43% and attributable mortality 9.5%. These findings compare
favorably with current reports of historical outcomes of 54, 67 and 28%, respectively (Caillot et al., 2001; Kim et al., 2001). A linked approach of early and serial CT scanning with moderately high doses of liposomal amphotericin B in neutropenic patients at high risk for IA improves the therapeutic outcome. The additional finding that the plain chest radiographs were either normal or non-specific at the time of the positive CT scans reinforces the need to optimize radiologic investigation. 3.3. Subcellular components Precision diagnostics are currently focused on the detection of fungal cell wall constituents and their antibodies and fungal nucleic acid in host body fluids. The PCR is oversensitive because of colonization-contamination but the use of a competitive PCR which raises the threshold for detection more closely relates to increased fungal load, reducing less false positives. Using universal primers to target conserved RNA gene sequences in all fungi—a panfungal PCR—is followed by exploiting the highly variable internal transcribed spacer regions through nested PCR to determine precise species (Sandhu et al., 1995). Cell wall antigen detection, e.g. galactomannan in Aspergillus is also oversensitive because of cross-reacting epitopes. However, it can be found early, before symptoms and signs, and titers fall if treatment is successful. Galactomannan may appear before PCR. In patients who survive, galactomannan levels fall in stark contrast to patients who die who have unchanged high levels (Maertens et al., 2001). Sensitivities and specificities for Aspergillus diagnosis are among the highest when a combination of ELIZA and PCR is used (Bretagne et al., 1998). The galactomannan sandwich ELIZA has had its credibility validated against histology/culture based gold standard
M. Ellis / Molecular Immunology 38 (2001) 947–957
reference points as a highly sensitive (90%) and specific (98%) non-invasive diagnostic tool recently. A potential benefit of successful application of this diagnostic test would be to rationalize the use of broad spectrum antifungals. It might be possible to reduce their use by 75% compared to use classical clinical diagnostic criteria. Cost, toxicity, emergent resistance could all be favorably influenced by selectively using antifungal drugs, based on positive serologic findings. However, a negative result cannot exclude Fusarium spp., Alternaria spp. and other molds—and so one could not withhold fungal treatment on the basis of a negative ELIZA result alone. A number of other issues need clarification. For example, there is a high false positive rate in the galactomannan assays—around 15%—the reason for this remains unclear. Dietary galactomannan translocation across damaged mucosa, cross-reactivity with Penicillium spp. and other antigens and partial treatment responses are all possible explanations. The other areas for investigation include optimal sampling frequency, the significance of galactomannan levels at different stages of infection and its usefulness to monitor and guide therapy.
4. Treatment of IFI 4.1. Antifungals The priority to define and use an antifungal agent having a substantial margin of safety is underscored from the adverse
951
experience with conventional amphotericin B (CAB), the well tried gold standard of care. Systemic toxicity occurs in 80% of patients. If CAB related renal failure occurs, the mortality in cancer patients triples to 54%, length of hospital stay increases to 8 days and there is a cost increment increase of US$ 30,000 per patient to treat the renal failure (Bates et al., 2001). On the other hand, attempts to reduce the incidence and severity of both general systemic reactions and local target organ toxicity, particularly renal, have in general been disappointing. Saline loading for example may not be tolerated in patients with hematologic malignancy, many of whom have cardiovascular compromization or increased pulmonary capillary permeability. Steroids and opiates, to reduce chills and rigors are used in around 50% of patients (Gallis et al., 1990). Their administration exposes the patient from further drug adverse reactions, which include heightening the immunocompromization status. Simple measures such as slowing the infusion rate may be beneficial in reducing severe chills and rigors. In a prospective double blind study of CAB given either over 1 or 4 h, the incidence of severe chills and rigors and tachycardia were significantly less with the slower infusion (Ellis et al., 1992; Fig. 4). Recently this has been extended to a 24 h continuous infusion (Eriksson et al., 2001). One of the postulated mechanisms for renal toxicity is CAB induced glomerular toxicity. Low dose dopamine has been investigated for its potential to prevent such a side effect, utilizing its physiologic property of increasing glomerular perfusion.
Fig. 4. Mean maximum pulse rise in two infusion rates of amphotericin B.
952
M. Ellis / Molecular Immunology 38 (2001) 947–957
Disappointingly a controlled study not only failed to show a significant protective effect but there was a high incidence of dopamine related adverse reactions (Camp et al., 1998). One of the most notable advances has been the reclothing of this emperor of antifungal therapy. Lipidisation had as its goal, an increase in the therapeutic index as a result of reduction in the toxic denominator and targeting of the AF at the site of fungal sepsis. Clearly this is an important strategic concept as there is considerable opportunity to correct the subtherapeutics arising from underdosing, dose interruption/discontinuation and inappropriate steroid, opiate and dopamine use. There are three commercially available lipid formulations of amphotericin B: ABLC, ABCD and LAB. These three differ considerably in structure, pharmacokinetics and pharmacodynamics. These aspects have been reviewed in detail elsewhere (Arikan and Rex, 2001). Each of the three compounds is unique in its properties; this is reflected in the difference in toxicity between them. The liposomal preparation (LAB) currently appears to have documented superior performance in this respect. A randomized study by Wingard et al. (2000) showed a significant less incidence in chills, rigors, nephrotoxicity and therapy discontinuations with LAB compared to ABLC. The efficacy impact of these highly expensive compounds has been the subject of intense debate (Ellis, 2000). Dose for dose, animal models demonstrate they have better mycologic eradication and survival compared to CAB. For example, there is less hematogenous dissemination from primary lung foci, increased lung sterilization and improved survival when the lipid compounds are compared to CAB (Leenders et al., 1996; Allende et al., 1994). A vast amount of information is also provided by open label human studies which include salvage treatment situations. These confirm high complete response rates with cure rates of generally around 60–70% for IA when various lipid formulations are used (Ellis, 2000). Controlled human clinical trials confirm the heightened efficacy of LAB over CAB. For example, there is improved fever resolution in patients with ARNF (Prentice et al., 1997), reduction in breakthrough IFI in patients at risk (Walsh et al., 1999) and a better response in the treatment of patients with established IFI (Leenders et al., 1998). Despite over 40 years of experience with amphotericin B there is no clear evidence which defines an optimal dose. Careful scrutiny of results from animal models reveals an interesting phenomenon. Mycologic eradication is clearly better at higher dosages, but survivorship does not parallel this to anything like the same degree— “mycologic-survival disassociation” (Ellis, 2000). This is clearly seen in Allende’s study (1994) where survivorship in the granulocytopenic rabbit infected with Aspergillus is 60% for 1 and 5 mg/kg/day of ABCD and 25% for 10 mg. Further support for this has been forthcoming in the only randomized controlled clinical trial of antifungal dosing
ever to be performed. EORTC #19923 showed that 1 and 4 mg of LAB in treating probable and proven IA were equi-effective in response and survival (Ellis et al., 1998). This apparently surprising finding has prompted many questions. The fact that a small group of patients with biopsy proven IA tended to do better with the higher dose suggested the trial may not have been powered sufficiently to show any significant difference at the higher dose. Perhaps a much higher dose differential was needed to be effective. Cenciu’s (1997) animal model in which a maximum tolerated dose was 100% successful suggests that might be the case. The ability of the controlled clinical trial as a model for investigating antifungal efficacy has recently been questioned (Rex et al., 2001). Patient accrual is generally poor for reasons which include poor patient performance status, reluctance of physicians to depart from time-honored antifungal treatments and an increasing number of new antifungals which seriously dilute trial patient base numbers. There are also many confounding variables in these very ill patients, which could influence outcome measures and which are difficult to control. Finally there is a widespread high degree of diagnostic variability—one review of over 170 studies gave a low κ score in this regard (Ascioglu et al., 2001). Carefully performed in depth case studies will be helpful in solving some problems such as dosing issues. For example, Patterson (2000) showed that in a patient that died from IPA, the antifungal drug concentrations were all low and below the MICs in the infected lung tissue but above the MICs in the adjacent healthy tissue. This suggested that poor outcome in patients with invasive mycoses may be in part related to failure to effectively deliver antifungals to the site of infection. 4.2. Recent antimycotic drugs A plethora of AF drugs with cell wall and subcellular targets other than the cell membrane have appeared in recent years. There has been a radical change in philosophy from “find an antifungal and then find a fungus to treat” to “investigate the genome for virulence factors and synthesize an antifungal to target those”. Among the third generation azoles is voriconazole whose target is the 14-␣-demethylase enzyme responsible for ergosterol synthesis of the cell wall, with remarkable broad spectrum activity to yeasts and molds. Posaconazole is another new generation azole with substantial (75% response in salvage therapy) effect against one of the most pan-resistant fungi—Fusarium spp. The candins which inhibit 1,3--glucan are remarkable for lack of any cross-resistance with other antifungals. Among these is capsofungin, recently licensed for the treatment of refractory IA. Others include the pradimicins which bind to mannosides and disrupt fungal cell wall activity, nikkomycin chitin inhibitors and the sordarins which inhibit fungal elongation factor 2.
M. Ellis / Molecular Immunology 38 (2001) 947–957
4.3. Exploitation of existing antifungals Molecular-stearic remodeling has already been exemplified with the lipid formulations of CAB. Nystatin has also been reformulated as a liposomal preparation, with major toxicity reduction. Liposomal forms of miconazole and ketoconazole are also under investigation. Recently immunoliposomes have been created. These contain fungus specific antibodies to facilitate directed fungal cell targeting. Complexing lipsosomal amphotericin B with polyethylene glycol (PEG-LAB) substantially increases the circulating half life and target fungi for longer periods of time. However, these preparations appear to yield less intracellular concentration of amphotericin B compared with the parent CAB, which may translate into inferior efficacy in the long term. Cochleates, lipid based supramolecular assemblies containing negatively charged phospholipid, natural products and divalent cations are among the latest innovative formulation vehicles to undergo clinical testing. Combination therapy of two different antifungals in general has not produced clear clinical advantages, apart from either fluconazole or amphotericin B in combination with 5-fluorocytosine which have synergistic activity against Cryptococcus neoformans meningitis. Combination of a primary antifungal with an agent having a different action has produced some important observations. For example, an important mechanism in Candida resistance is efflux pumping of azoles out of the yeast cell prior to target binding. White (1997) demonstrated in HIV patients that as the duration of the use of fluconazole increased over several months, mRNA for the efflux pump was progressively expressed which paralleled the rise in MICs and use of higher fluconazole doses. The use of MC-510027 an efflux pump inhibitor has now been shown to reduce the MICs of the azoles in such situations (Chamberland et al., 1999).
5. Experimental strategies It is crucial that therapeutic modalities are explored which have interventional characteristics other than direct antimicrobial. Fungal resistance, clinical response failures in the presence of fully sensitive organisms, overwhelming acute onset septic states that occur before an antifungal can be reasonably expected to work—are some of the compelling reasons. Current experimental approaches include the following. 5.1. Hyperbaric oxygen (HBO) This unusual approach harvests the demonstrable properties of oxygen under increased tension such as augmentation of white cell mediated fungal killing, optimization of tissue pO2 and reduction of edema and acidosis. Substantial improvements in outcome have been documented for
953
selected fungal infections such as mucormycosis (Price and Stevens, 1980). Fig. 5 shows a patient with hematogenous IA whose large necrotic skin lesions did not respond to LAB until hyperbaric oxygen (HBO) was added as adjunctive therapy. 5.2. Surgical adjunctive therapy The surgical management of IFI aims to substantially reduce large fungal masses or to remove dormant lesions which could re-activate during future immunosuppression. In patients with hematologic malignancy and IPA, resection of well circumscribed lung masses during the post neutrophil recovery phase reduces future chemotherapy induced reactivation of the IPA by 50% (Karp et al., 1988). Mortality from catastrophic bleeding associated with the mycotic lung sequestrum of lung cavitary aspergillosis can also be reduced by timely surgical intervention in these patients (Wong et al., 1992). Fungal endocarditis is another illustration of the mandatory use of adjunctive surgical intervention. A meta-analysis of the world literature indicated that outcome in such patients was improved when antifungal treatment was combined with surgical valve replacement (Ellis et al., 2001). This approach appears to be the most beneficial for patients with Aspergillus endocarditis as mortality is virtually 100% with amphotericin B treatment alone. 5.3. Targeting virulence-pathogenic factors The phospholipases of C. albicans were first recognized 40 years ago. These are produced in great quantities shed from the very tips of the hyphae and are correlated with tissue invasion and cell cytolyis. Parallelism exists between phospholipase production, hematogenous dissemination and mortality (Ghannoum, 2000). At least three genes have been identified. Not only do the phospholipases permit host cell entry, but they deregulate cell signaling, induce IL-8/-6 expression and induce inflammatory mediators such as aracidonic acid. Synthetic phospholipase inhibitors have been manufactured which prolong animal survival when used with fluconazole to treat severe sepsis (Hanel et al., 1995). 5.4. Neutrophil properties One of the pivotal requirements for recovery from IFI is correction of neutropenia. The normal observation in a patient with IA and neutropenia is that recovery of neutropenia is mandatory for effective recovery from the IA, despite antifungal agents. Previous concerns that white cell transfusions posed major problems of safety and efficacy have been largely quelled with current state of the art cell harvest technology, donor screening for blood borne infections and assurance of good quality and quantity of white cells made possible by the use of recombinant colony stimulating factors. There is resurrected interest in granulocyte transfusions to support such patients with severe IFI (Price et al., 2000).
954
M. Ellis / Molecular Immunology 38 (2001) 947–957
Fig. 5. (a) Cutaneous hematogenous aspergillosis before HBO treatment: note ischemia and edema; (b) same patient after HBO treatment: note resolution of signs and healthy granulation tissue.
It is the primary azurophil granules and the secondary granules of the neutrophil that contain a variety of naturally occurring antimicrobial peptides and proteins and contribute towards the host’s innate immune response. These are deployed at the sites of fungal infection. Seroprocidin in one example—it is in fact a family which produce lethal membrane perturbations (Levy, 2000). The possibility of utilizing recombinant preparations is now a reality facing clinical testing. 5.5. Soluble immune system enhancement This appears to be of crucial importance in the successful response to IFI in the immunocompromized. Unfortunately, the rapidity with which colony stimulating factors were introduced into clinical practice and are now widely
used almost on a routine basis precludes a pure evidence based evaluation of their efficacy. There is good circumstantial evidence however for an improved outcome when granulocyte, monocyte growth factors, interferon-␥, IL-12, TNF-␣ and others are used as adjunctive therapy. The ECOG study of Rowe (1998) showed significant improved survivorship when colony stimulating factors were used in addition to antifungal treatment of IFI. Immunomanipulation may also be beneficial in the “not so immunocompromized patient” with an invasive mycosis. An example of this phenomenon is invasive sinus aspergillosis usually due to Aspergillus flavus, characterized by progressive erosive and destructive skull base disease and intracranial spread from an ethmoid sinus epicenter (deShazo et al., 1997). Many such patients have no gross or apparent immunodeficiency. One such patient had subtle immune
M. Ellis / Molecular Immunology 38 (2001) 947–957
955
Fig. 6. (a) Extensive intracranial aspergillosis, spreading through cribriform plate (MRI SE T1 weighted image); (b) same patient after combination antifungal and immunomodulatort treatment showing extensive disease reduction.
defects noted (Ellis et al., 2002). These were: CD4-penia (458 cells/l), low percentages of CD3+ , CD15+ , and natural killer cells (5%), weak quantitative nitroblue tetrazolium reduction responses (31–50% normal) and high IgE levels (433 IU/ml) in the absence of allergy. Taken together these findings suggested a selective Th1 cell defect, which could have predisposed the patient to invasive aspergillosis. She presented with a massive intracranial A. flavus infection which included MRI-scan documentation of infiltration of the optic nerves and visual loss. Despite up to 10 mg/kg/day of liposomal amphotericin B she deteriorated. She did not respond until GMCSF and ␥-interferon were introduced as part of the therapy, consequently clearing most of the intracerebral fungal load (Fig. 6). GMCSF is known to enhance neutrophil, monocyte and macrophage antifungal activities in vivo, and ␥-interferon improves antifungal properties of polymorphonuclear cells and of helper T cells by directing their activities towards a Th1 response and inducing neutrophil mediated fungal hyphal damage. The case study underscores on the one hand our limited understanding of the precise fungal–host interactive mechanisms in such patients and on the other, despite this, such patients appear to respond to combination antifungal-immunomodulatory treatments.
Further clarification in this field will define subpopulations more likely to benefit from such strategies. 5.6. Prophylactic cytokines Finally, a recent interesting concept is that of the cytokine shield. The majority of invasive yeasts and bacteria arise from the gut, translocating as a result of chemotherapy induced apoptosis of the gut epithelium with consequent pathologica changes in gut wall permeability. Recombinant IL-11 used for managing thrombocytopenia also has epithelial growth potential. The hypothesis is that if gut integrity can be preserved through the chemotherapy treatment phase, fungal and bacterial infections might be reduced in frequency and severity. A neutropenic mouse model showed that survival following a lethal dose of Pseudomonas aeruginosa could be increased with r IL-11 (Opal et al., 1999). A double blind placebo controlled trial of recombinant IL-11 given to leukemic patients to assess its effect on preventing sepsis is underway. An interim safety analysis of the first 20 patients, whilst preserving anonymity of the trial drugs has indicated a difference in some of the end points of the two study arms. Septic events, fungal events, severity of sepsis
956
M. Ellis / Molecular Immunology 38 (2001) 947–957
Table 2 Recombinant interleukin-11 in leukemic patients Preliminary crude data on 20 patients
Patients (n) Total infection events (n) Sepsis score ≥ 3a Any bacteremiaa Cryptic fungal infectionsa Abnormal lactulose: mannitol ratioa Lactulose:mannitol ratio (median) Adverse eventsa
Drug A
Drug B
11 11 3 8 6 4b
9 7 1 4 2 2c
0.12
0.02
3 (injection pain, injection pain, edema)
2 (edema, pleural effusion)
Normal lactulose:mannitol ratio < 0.01. a Number of patients. b Associated with bacteremias due to Pseudomonas, Enterococcus, Staphylococcus, Klebsiella, and Fusobacterium spp. c Associated with bacteremia due to Klebsiella spp.
and intestinal epithelial cell integrity (as assessed with urinary lactulose:mannitol ratios) were all consistently different between the two arms (Table 2). Should the ultimate findings suggest an advantage for recombinant IL-11 in reducing fungal and bacterial sepsis, this would give further encouragement for exploration of such non-antimicrobial based strategies in managing invasive mycoses. References Abi-Said, D., Anaissie, E., Uzun, O., et al., 1997. The epidemiology of hematogenous candidiasis caused by different Candida species. CID 24, 1122–1128. Allende, M.C., Lee, J.W., Francis, P., et al., 1994. Dose dependent antifungal activity and nephrotoxicity of amphotericin B colloidal dispersion in experimental pulmonary aspergillosis. Antimicrob. Agents Chemother. 38, 518–522. Arikan, S., Rex, J., 2001. Lipid-based antifungal agents: current status. Curr. Pharma. Des. 7, 393–415. Ascioglu, S., de Pauw, B., Donnely, J., et al., 2001. Reliability of clinical research on invasive fungal infections: a systematic review of the literature. Med. Mycol. 39, 35–40. Bates, D., Su, L., Yu, D., et al., 2001. Mortality and costs of acute renal failure associated with amphotericin B therapy. CID 32, 686–693. Bretagne, S., Costa, J., Bart-Delabesse, E., et al., 1998. Comparison of serum galactomannom antigen detection and competitive polymerase chain reaction for diagnosing invasive aspergillosis. CID 26, 1407– 1412. Caillot, D., Casasnovas, O., Bernard, A., et al., 1997. Improved management of invasive pulmonary aspergillosis in neutropenic patients using early thoracic computed tomographic scan and surgery. J. Clin. Oncol. 15, 139–147. Caillot, D., Couaillier, J., Bernard, A., et al., 2001. Increasing volume and changing characteristics of invasive pulmonary aspergillosis on sequential thoracic computer tomography scans in patients with neutropenia. J. Clin. Oncol. 19, 253–259. Camp, M., Wingard, J., Gilmore, C., et al., 1998. Efficacy of lowdose dopamine in preventing amphotericin B nephrotoxicity in bone marrow transplant patients and leukemia patients. Antimicrob. Agents Chemother. 42, 3103–3106.
Cenciu, E., Perito, S., Enssle, K., et al., 1997. Th1 and Th2 cytokines in mice with invasive aspergillosis. Infect. Immunol. 65, 564–570. Chamberland, S., Blais, J., Cotter, D., 1999. Impact of MC-510,027: a fungal efflux pump inhibitor, on the susceptibility of clinical isolates of Candida spp. to antifungal agents. In: Proceedings of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, September 1999. Abstract #1270. deShazo, R., O’Brier, M., Chapin, K., et al., 1997. A new classification and diagnostic criteria for invasive fungae sinusitis. Arch. Otolarygol. Head Neck Sing. 123, 1181–1188. Ellis, M., 2000. Amphotericin B preparations: a maximum tolerated does in severe invasive fungal infections? Transplant Infect. Dis. 2, 51–61. Ellis, M., al-Hokail, A., Clink, H., et al., 1992. Double blind randomized study of the effect of infusion rates on toxicity of amphotericin B. Antimicrob. Agents Chemother. 36, 172–179. Ellis, M., Qadri, S., Spence, D., et al., 1994. The effect of fluconazole as prophylaxis for neutropenic patients on the isolation of Candida species from surveillance cultures. J. Antimicrob. Chemother. 33, 1223–1228. Ellis, M., Spence, D., Bouchama, A., et al., 1995. Open lung biopsy provides a higher and more specific diagnostic yield compared to broncho-alveolar lavage in immunocompromised patients. Scand. J. Infect. Dis. 27, 157–162. Ellis, M., Spence, D., de Pauw, B., et al., 1998. An EORTC international multicenter randomized trial (EORTC number 19923) comparing two dosages of liposomal amphotericin B for treatment of invasive aspergillosis. CID 27, 1406–1412. Ellis, M., Al-Abdely, H., Sandridge, A., et al., 2001. Fungal endocarditis: evidence in the world literature 1965–1995. CID 32, 50–62. Ellis, M., Watson, R., McNabb, A., et al.., 2002. Massive intracerebral aspergillosis responding to combination high dose liposomal amphoetricin B and cytokine therapy without surgery. J. Med. Microbiol. 51, 70–75. EORTC, International Antimicrobial Therapy Cooperative Group, 1989. Empirical antifungal therapy in febrile granulocytopenic patients. Am. J. Med. 86, 668–672. Eriksson, U., Seifert, B., Scaffner, A., 2001. Comparison of effects of amphotericin B deoxycholate infused over 4 or 24 h: randomized controlled trial. BMJ 322, 579–582. Gallis, H., Drew, R., Pickard, W., 1990. Amphotericin B: 30 years of clinical experience. Rev. Infect. Dis. 12, 308–329. Gerson, S., Talbot, G., Hurwitz, S., et al., 1985. Discriminant scorecard for diagnosis of invasive pulmonary aspergillosis in patients with acute leukemia. Am. J. Med. 79, 57–64. Ghannoum, M., 2000. Potential role of phospholipases in virulence and fungal pathogenesis. Clin. Microbiol. Rev. 13, 122–143. Groll, A., Shah, P., Mentzel, C., et al., 1996. Trends in the postmortem epidemiology of invasive fungal infections at a university hospital. J. Infect. 33, 23–32. Gunsilius, E., Lass-Florl, C., Kahler, C., 2001. Candida ciferrei, a new fluconazole-resistant yeast causing systemic mycosis in immunocompromized patients. Ann. Hematol. 80, 178–179. Hanel, H., Kirsch, R., Schmidts, H., et al., 1995. New systematically active antimycotics from the -blocker category. Mycoses 38, 251–264. Jarvis, W., 1995. Epidemiology of nosocomial fungal infections, with emphasis on Candida species. CID 20, 1526–1530. Karp, J., Burch, P., Merz, W., 1988. An approach to intensive antileukemia therapy in patients with previous aspergillosis. Am. J. Med. 85, 203– 206. Kim, M., Lee, K., Kim, J., et al., 2001. Crescent sign in invasive pulmonary aspergillosis: frequency and related CT and clinical factors. J. Comput. Assist. Tomogr. 25, 305–310. Levy, O., 2000. Antimicrobial proteins and peptides of blood: templates for novel antimicrobial agents. Blood 96, 2664–2672. Leenders, A.C.A.P., De Marie, S., Ten Kate, M.T., Bakker-Woudenberg, I.A.J.M., Verbrugh, H.A., 1996. Liposomal amphotericin B (Ambisome) reduces dissemination of infection as compared with amphotericin B deoxycholate (Fungizone) in a rat model of pulmonary aspergillosis. J. Antimicrob. Chemother. 38, 215–225.
M. Ellis / Molecular Immunology 38 (2001) 947–957 Leenders, A.C.A.P., Daenen, S., Jansen, R.L.H., et al., 1998. Liposomal amphotericin B compared with amphotericin B deoxycholate in the treatment of documented and suspected neutropenia associated invasive fungal infections. Br. J. Haematol. 103, 205–212. Maertens, J., Verhaegen, J., Lagrou, K., et al., 2001. Screening for circulating galactomannan as a noninvasive diagnostic tool for invasive aspergillosis in prolonged neutropenic patients and stem cell transplantation recipients: a prospective evaluation. Blood 97, 1604– 1610. Opal, S.M., Jhung, J.W., Keith, J.C., et al., 1999. Additive effects of human recombinantinterleukin-11 and granulocyte colony-stimulating factor in experimental gram-negative sepsis. Blood 93, 3467–3472. Patterson, P., Johnson, B., Ainscough, S., et al., 2000. Treatment failure in invasive aspergillosis due to poor tissue penetration by antifungal agents. In: Proceedings of 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, September 2000. Abstract #1328. Pfaller, M., Jones, R., Doern, G., et al., 1998a. International surveillance of bloodstream infections due to Candida species: frequency of occurrence and antifungal susceptibilities of isolates collected in 1997 in the United States, Canada and South America for the SENTRY programme. The SENTRY participant group. J. Clin. Microbiol. 36, 1886– 1889. Pfaller, M., Jones, R., Messer, S., et al., 1998b. National surveillance of nosocomial blood stream infection due to Candida albicans: frequency of occurrence and antifungal susceptibility in the SCOPE program. Diagn. Microbiol. Infect. Dis. 31, 327–332. Powderley, W., Finkelstein, D., Feinberg, J., et al., 1995. A randomized trial comparing fluconazole with clotrimazole troches for the prevention of fungal infections in patients with advanced human immunodeficiency virus infection. NEJM 332, 700–705. Prentice, H.G., Hann, I.M., Herbrecht, R., et al., 1997. A randomised comparison of liposomal versus conventional amphotericin B for the treatment of pyrexia of unknown origin in neutropenic patients. Br. J. Haematol. 98, 711–718. Price, J., Stevens, D., 1980. Hyperbaric oxygen in the treatment of rhinocerebral mucormycosis. Laryngoscope 90, 737–747.
957
Price, T., Bowden, R., Boeckh, M., et al., 2000. Phase I/II trial of neutrophil transfusions from donors stimulated with G-CSF and dexamethasone for treatment of patients with infections in hematopoietic stem cell transplantation. Blood 95, 3302–3310. Rees, J.R., Pinner, R., Hajjeh, R., et al., 1998. The epidemiological features of invasive mycotic infections in the San Francisco Bay area, 1992–1993: results of population-based laboratory active surveillance. CID 27, 1138–1147. Rex, J., Walsh, T., Nettleman, M., et al., 2001. Need for alternative trial designs and evaluation strategies for therapeutic studies on invasive mycoses. CID 33, 95–106. Rowe, J., 1998. Treatment of acute myeloid leukemia with cytokines: effect on duration of neutropenia and response to infections. CID 26, 1290–1294. Sandhu, G., Kline, B., Stockman, L., et al., 1995. Molecular probes for diagnosis of fungal infections. J. Clin. Microbiol. 33, 2913–2919. Verweij, P., Donnelly, J., de Pauw, B., et al., 1996. Prospects for the early diagnosis of invasive aspergillosis in the immunocompromised patient. Rev. Med. Microbiol. l7, 105–113. Viscoli, C., Girmenia, C., Marinus, A., et al., 1999. Candidemia in cancer patients: a prospective, multicenter surveillance study by the Invasive Fungal Infection Group (IFIG) of the European Organisation for Research and Treatment of Cancer (EORTC). CID 28, 1071–1079. Walsh, T.J., Finberg, R.W., Arndt, C., et al., 1999. Liposomal amphotericin B for empirical therapy in patients with persistent fever and neutropenia. N. Engl. J. Med. 340, 764–771. White, T., 1997. Increased mRNA levels of ERG16, CDR, and MDR1correlate with increases in azole resistance in Candida albicans isolates from a patient infected with human immunodeficiency virus. Antimicrob. Agents Chemother. 41, 1482–1487. Wingard, J., White, M., Anaissie, E., et al., 2000. A randomized double-blind comparative trial evaluating the safety of liposomal amphotericin B versus amphotericin B lipid complex in the empirical treatment of febrile neutropenia. CID 31, 1155–1163. Wong, K., Waters, C., Walesby, R., 1992. Surgical management of invasive pulmonary aspergillosis in immunocompromised patients. Eur. J. Cardio-Thorac. Surg. 6, 138–143.