Aspergillus flavus: an emerging non-fumigatus Aspergillus species of significance

Summary

Invasive aspergillosis is rare in immunocompetent people but contributes to significant morbidity and mortality in immunosuppressed patients. The majority (approximately 80%) of invasive Aspergillus infections is caused by Aspergillus fumigatus. The second most frequent (approximately 15–20%) pathogenic species is Aspergillus flavus and to a lesser extent, Aspergillus niger and Aspergillus terreus. Aspergillus flavus has emerged as a predominant pathogen in patients with fungal sinusitis and fungal keratitis in several institutions worldwide. To date, there has not been any publication exclusively reviewing the topic of A. flavus in the literature. This article reviews the microbiology, toxigenicity and epidemiology of A. flavus as well as describes the clinical characteristics, diagnosis and management of infections caused by this organism.

Introduction

Members of the genus Aspergillus are ubiquitous moulds widely distributed in the environment. About 185 different species of Aspergillus have been identified, of which 20 are documented to cause human disease. Aspergillus spores, upon inhalation, can lead to colonisation, allergic manifestations or invasive infection depending on host immunity. Invasive aspergillosis is the second most common invasive fungal infection in humans in the United States, next only to candidiasis. Invasive aspergillosis is rare in immunocompetent people¹ but contributes to significant morbidity and mortality in immunosuppressed patients. Majority (approximately 80%) of invasive infections is caused by Aspergillus fumigatus. The second most frequent (approximately 15–20%) pathogenic species is Aspergillus flavus² and to a lesser extent, Aspergillus niger and Aspergillus terreus. Although A. fumigatus is the most common aetiological agent of invasive aspergillosis overall in the United States, A. flavus, with its unique ability to survive at higher temperatures, is the predominant pathogen in countries with arid dry conditions, including most of the Middle East, Africa and Southeast Asia accounting for 50–80% of cases of sinusitis, keratitis and cutaneous infections in humans (Table 1).^2–10 There have been two published reviews of childhood aspergillosis so far, one from Toronto (1979–1988) and the other from St Jude’s Children’s Hospital in Memphis (1962–1996). The predominant species isolated from both these studies was A. flavus with an incidence of 65% and 72% respectively.^11,12 Also, A. flavus is a unique species of Aspergillus, in that it has a wide range of hosts including humans, animals and plants and is the predominant pathogen involved in the production of aflatoxins.

Table 1.   Morphological and epidemiological characteristics of Aspergillus flavus and A. fumigatus

Characteristics	A. flavus	A. fumigatus
Sclerotia in tissue	Present	Absent
Toxin production	Aflatoxin production in tissue	No aflatoxin production
Hyphal growth	Relatively slow	Fast
Pulmonary	Less common, allergic forms not reported	More common, allergic forms are exclusively caused by A. fumigatus
ENT infections, eye	More common in Sudan, India, Sri Lanka, Saudi Arabia, few institutions in USA	Emerging in the US
Cutaneous infections	Usually primary, at central venous catheter sites	Usually secondary from dissemination
Solitary emboli	Reported	No data
Immune status	Case reports in immunocompetent	Lower incidence in immunocompetent
Diagnosis	Higher yield with tissue rather than bronchoalveolar lavage/sputum^197 Has relatively high MICs	MICs lower than that for A. flavus
Resistance in vitro	Rapid selection of resistant mutants (laboratory observation, unpublished)	Requires repeated subcultures at higher drug concentrations

We present a brief overview of the unique characteristics of this potentially pathogenic dimorphic fungus and its clinical relevance as an emerging pathogen.

Morphological characteristics

Aspergillus spp. belong to Ascomycota phyllum and Trichomaceae family. Species are differentiated by morphological characteristics and colour. Macroscopically, colonies are flat, granular, downy to powdery in texture often with radial grooves. Colony surface is yellow initially but turns dark yellowish green with age, characteristically different from that of other commonly found Aspergillus spp., such as A. fumigatus, A. niger, A. nidulans and A. terreus. A. flavus hyphae are septate and hyaline branching at 45° angle. The conidiophores originate from supporting hyphae and terminate in vesicles at the apex. The conidiophores of A. flavus (up to 800 μm long and 15–20 μm wide) are rough and colourless and the vesicle (20–45 μm) is globose with radiations. The phialides may be uniseriate and attached to the vesicle directly or are biseriate and are attached to the vesicle via a supporting cell called metula (8 × 5 μm). The conidial heads are typically radiate, 400 nm in diameter, splitting to form loose columns. The conidia are 2–5 μm in diameter, are arranged in a radial chain over the distal end of the phialides. Microscopical structures called sclerotia can be identified in A. flavus and are of key importance in their identification.¹³In vitro, A. flavus grows well on Sabouraud dextrose agar or Czapek Dox and malt extract at 37 °C and germination of conidia occurs at about 24 h. Primary growth is by apical hyphal extension which may be relatively slower than that of A. fumigatus. Other growth characteristics are similar to A. fumigatus and growth reaches a plateau after 24 h (laboratory observation). Although the rate of hyphal extension has been associated with aggressiveness of A. fumigatus, comparable studies have not been performed for A. flavus.

Epidemiology

Aspergillus flavus has a worldwide distribution in the environment. Although it is universally found in air, soil, dust and water, higher fungal burdens have been noted particularly in contaminated peanuts, corn, grains and decaying organic matter.¹⁴ Disease prevalence is highly variable with different institutions reporting different species of Aspergillus as the predominant pathogen.² The primary mode of transmission to humans is by inhalation of conidia. There is increasing concern about contaminated food, environmental and occupational exposure to fungal spores of different species, especially to the aflatoxin-producing strains of A. flavus, in different parts of the world.^15–19 Higher frequency of pulmonary function impairment and allergic respiratory diseases including asthma has been reported in farmers around the world.²⁰ In addition to inhalation, a secondary route of transmission has been reported via contact with skin or wound (trauma and postoperative), contamination of intravenous (i.v.) solutions, wound dressings and marijuana inhalation.^21–24

Pathogenesis

Although unequivocal experimental evidence is lacking at present, many putative virulence factors are implicated by extrapolation from studies conducted in A. fumigatus for the infection and pathogenesis of A. flavus. Conidia (2–5 μm) that gain access to the pulmonary alveoli germinate into actively growing hyphae resulting in pulmonary infection as well as invasive disease. Several biochemical and cellular factors (mostly proteinases, the role of which is discussed in the following paragraph) produced by the infecting organism may aid the germination of conidia and the penetration of the emerging hyphae into the lung alveoli cells.

The roles of extracellular proteinases^25–27 such as secreted aspartyl proteinase (SAP), serine proteinase (SP), metalloproteinase (MP), alkaline proteinase and lipases^28,29 have been studied extensively during infection and pathogenesis of Candida species. Cellular, biochemical and molecular studies have indicated that one or more of these extracellular enzymes act as virulence factor(s) for the infection. Many investigators believe that a similar situation prevails in the case of pathogenic filamentous fungi, in particular for Aspergillus spp., including A. flavus.^30–36 However, it is possible that the role of extracellular proteinases could be twofold. As these organisms, in particular Aspergillus spp. are commonly found in niches where the availability of nutrients is very scarce, they have to utilise substrates available as nutrients. The extracellular proteinases produced by these organisms help to breakdown complex protein molecules to utilisable nutrients. In addition, these enzymes digest the extracellular matrix made up of elastin and collagen fibres that facilitates tissue penetration.

The SAP, SP and MP are the most commonly found pathogen-produced extracellular proteinases associated with the infection of Aspergillus spp., including A. flavus. These extracellular proteinases are called thus because of the presence of an aspartic acid residue or a serine residue or a metal ion (e.g. zinc), respectively, at the active sites of these enzymes. Although the contribution of SAPs for the infection attributable to A. flavus per se is lacking, it is to be noted that several groups have investigated the possible involvement of SAP during the infection and pathogenesis of other animal and plant filamentous fungi^37,38 such as Aspergillus oryzae³⁹ and Aspergillus nidulans,⁴⁰ that are non-pathogenic in humans, Botrytis cinerea,⁴¹A. fumigatus,^42,43Rhizopus oryzae,⁴⁴Podospora anserine⁴⁵ and Mucor pusillus.⁴⁶ On the other hand, Reichard et al. [47] have investigated the possible role of SAP during the infection and pathogenesis of A. fumigatus using an A. fumigatus mutant deficient of aspergillopepsin, and they concluded that it is probable that the proteinases do not contribute decisively to tissue invasion in the pathogenesis of systemic aspergillosis.

Although the active participation of SAP in the infection and pathogenesis of A. flavus is only implicated by studies performed in other pathogenic filamentous fungi, several studies have indicated that SP is most probably a virulence factor for A. flavus,^30–36A. fumigatus^48–57 and A. terreus.⁵⁸ The effect of this enzyme as a virulence factor can be either direct or indirect, namely, acting on the extracellular matrix of the protective tissue(s) or by inhibiting the neutrophil function highly required for the clearance of fungal infection. In contrast to the positive correlation of the presence of SP and virulence of Aspergillus spp., including A. flavus, several investigators by immunological, genetic, biochemical techniques as well as animal models, have failed to show a positive correlation between the production of SP and pathogenesis, casting a shadow of suspicion regarding the role of this enzyme as a virulence factor.⁴⁷

The third class of extracellular proteinases known to augment infection and pathogenesis of A. fumigatus and A. flavus is the MP. They are called MPs because their active site often binds a metal ion (e.g. zinc ion) as a co-factor. Rhodes et al. [59,60] and Ramesh et al. [61] have identified an extracellular proteinase that fits the biochemical profile of a MP in A. flavus. Preliminary studies have implicated an active role for this enzyme for the virulence of A. flavus.

Some reports suggest that the disease process may be potentiated by aflatoxins,⁶² particularly in the immunocompromised/neutropenic host. If aflatoxins function as virulence factors, the action of this class of compounds will be indirect by inhibiting the neutrophil function of the host. As indicated elsewhere, without a properly functioning cellular immune system, the clearance of Aspergillus infection is extremely difficult giving the infecting pathogen an advantage.

Production of pigment(s) by various Aspergillus spp., including A. flavus is generally considered to be an indirect virulence factor, although the pigments by themselves may not potentiate tissue invasion or the disease process. The chromogenic pigments characteristically produced by Aspergillus spp., including A. flavus, protect conidia from environmental factors such as heat, UV irradiation, extreme pH etc., as well as ingestion by macrophages, thus providing a clear survival advantage within and without the host, thereby indirectly acting as a virulence factor.

As the survival of conidia in the host is a prerequisite for establishing disease, several investigators have examined the effect of pigmentation during the infection and pathogenesis of Aspergillus spp., in particular A. fumigatus using genetic, biochemical and morphological factors as well as using animal models.⁶³ Similar studies, if performed, would enhance our knowledge on virulence factors and pathogenesis in A. flavus that would have clinical relevance.

A. flavus and immune cell interactions

Research on the immuno-modulatory effects of various A. flavus antigens and aflatoxins in humans is still preliminary. Nevertheless, in vitro and in vivo studies published so far have reported that both surface-bound and exoantigens of A. flavus are capable of inducing specific monoclonal antibodies and cellular immunity. It appears from murine studies in vivo that aflatoxin B (AFB1) treatment preferentially affects macrophage functions, and in particular, it decouples the close correlation usually observed between transcriptional and translational controls of interleukin-1 (IL-1) alpha, IL-6 and tumour necrosis factor production by these cells. The results suggest that AFB1 had a direct and complex effect on lymphocytes and there is a differential sensitivity of various subpopulations. In addition, it has also been demonstrated that protein A may prove to be a useful agent to protect the host from aflatoxin immunotoxicity, in view of its stimulatory effects on various immune functions even after their initial depression attributable to aflatoxin B1.^12,64–75

Toxigenicity

Aspergillus flavus and A. parasiticus are known to be responsible for the production of one of the most potent cancer-causing groups of mycotoxins, the aflatoxins. Although these fungal species have similar geographical ranges, A. parasiticus is less widely distributed and is rarely present in Southeast Asia, whereas A. flavus is the most widely distributed fungus responsible for aflatoxin contamination of foodstuffs in the world.

Like most other mycotoxins, these compounds are produced as secondary metabolites and not known to have any specific role in the biology of these organisms. The four major types of aflatoxins are B1 (molecular weight 312.3 Da), B2 (molecular weight 314.3 Da), G1 (molecular weight 328.3 Da) and G2 (molecular weight 330.3 Da), based on their fluorescent colour when exposed to ultraviolet light (B = blue fluorescence; G = yellow-green fluorescence). Aflatoxin M1, which may be found in the absence of other aflatoxins, is a major metabolic hydroxylation product of aflatoxin B1. These compounds are slightly soluble in aqueous solutions, soluble in moderately polar organic solvents and insoluble in non-polar solvents. They are unstable when exposed to oxidising agents, ultraviolet light or solutions with a pH below 3 and above 10. These compounds are not destroyed under normal cooking conditions, but can be completely destroyed by autoclaving in the presence of ammonia or by treating with bleach.^76,77

Aflatoxins are known to cause liver cancer (hepatocellular carcinoma or primary liver-cell cancer) in humans. Early evidence for the carcinogenicity of aflatoxins in humans came from descriptive studies that correlated geographical variation in aflatoxin content of foods with geographic variation in the incidence of liver cancer.^77–79 For instance, studies in Uganda, Switzerland, Thailand, Kenya, Mozambique, Philippines and China demonstrated strong, significant positive correlation between estimated aflatoxin intake or aflatoxin levels in food samples and the incidence of liver cancer. Similarly, in the United States, a 10% increase in hepatocellular cancer was observed in the Southeast, where the estimated average daily intake of aflatoxin was high, compared with the Northwest, areas with low aflatoxin intake.⁸⁰

The aflatoxin aetiology of carcinogenesis in humans is supported by studies in animal models. Mice, rats and other experimental animals exposed to individual or mixture of aflatoxins by various routes such as oral, intraperitoneal and subcutaneous (s.c.) developed tumours at multiple tissue sites.^77,79

Studies on the mechanism of action of aflatoxins show that in humans and susceptible animal species, aflatoxin B1 is metabolised by cytochrome P-450 enzymes to aflatoxin-8,9-epoxide, a highly reactive form that binds to DNA and to albumin in the blood-forming adducts. The 8,9-epoxide metabolite can be detoxified through conjugation with the cellular thiol glutathione, mediated by the enzyme glutathione S-transferase (GST). Interestingly, the activity of GST is much higher by a factor of 3–5 in animal species that are resistant to aflatoxin carcinogenicity (e.g. mice) than in susceptible animal species (e.g. rats). Humans have lower GST activity than either mice or rats, suggesting that humans are less capable of detoxifying aflatoxin-8,9-epoxide. Aflatoxin causes genetic damage in bacteria, in cultured cells from humans and experimental animals exposed to aflatoxin in vivo. The types of genetic damage observed include formation of DNA adduct, gene mutations (G to T transversion), micronucleus formation indicating chromosome damage or loss, sister chromatid exchange and mitotic recombination.^80–84

The general population is exposed to aflatoxins primarily by eating contaminated food. In the developing world consumption of grains, cereals and pulses contaminated with aflatoxin(s) is a major health problem. Poor storage conditions of these staple food items in the presence of high humidity and temperature provides an ideal environment for the growth these organisms in contaminated storage facilities. Meat, eggs, milk and other edible products from animals that consume aflatoxin-contaminated feed are additional sources of potential exposure. Data on contamination of foods compiled in 1995 from 90 countries reported a median aflatoxin B1 level of 4 μg kg⁻¹ (ranging 0–30 μg kg⁻¹) and a median total aflatoxin level of 8 μg kg⁻¹ (ranging 0–50 μg kg⁻¹).⁷⁹ Nursing infants may be exposed to aflatoxins (e.g. M1, 0.02–1.8 μg l⁻¹; B1, as high as 8.2 μg l⁻¹) in breast milk. For example, aflatoxins were detected in 90 of 264 breast-milk samples collected from nursing mothers in Africa but were not detected in 120 samples collected from nursing mothers in Germany highlighting the endemic nature of problem.^85–87

Exposure to aflatoxins can be an occupational health hazard for farmers and other agricultural workers by inhalation of dust generated during the handling and processing of the contaminated crops and feeds. Of 45 animal-feed production plant workers in Denmark, seven had detectable levels of aflatoxin B1 in their blood after working for 4 weeks in the factory or unloading raw materials from ships.⁸⁸ Gosh et al. [89] reported detecting aflatoxins at concentrations ranging from 0.00002 to 0.0008 μg m⁻³ in respirable dust samples collected in workplace and storage areas of corn-processing plants in India. Similarly, Selim et al. [90] collected dust samples from 28 farms in the United States during harvest and unloading, animal feeding and bin cleaning, and found aflatoxin concentrations ranging from 0.00004 to 4.8 μg m⁻³. These and several other studies⁹¹ have found that exposure to dust contaminated with aflatoxins can be an occupational health hazard.

Clinical presentations of invasive disease

Bronchopulmonary infections

Aspergillus flavus is responsible for approximately 10% of bronchopulmonary infections.^92–95 The low incidence of A. flavus infections compared with A. fumigatus overall could be related to the either the species distribution in the environment or to host factors. An interesting study from Portugal evaluated various parameters associated with germination of different species of Aspergillus conidia; they found that the internal human milieu provided ideal conditions for the development of invasive disease by A. fumigatus but restricted the invasion of A. flavus and A. niger.⁹⁶ Pulmonary manifestations of A. flavus are not different from those caused by A. fumigatus and may take the form of aspergilloma, invasive pulmonary aspergillosis⁹⁷ or tracheobronchial aspergillosis. Interestingly, allergic forms of aspergillosis specifically caused by A. flavus (extrinsic allergic alveolitis and allergic bronchopulmonary aspergillosis) have not been reported in the literature. This may be related to inter-species variability in eliciting a hypersensitivity response in response to antigenic stimulation.^98–104 Asymptomatic aspergillomas secondary to A. flavus are rare; as with A. fumigatus, they require no therapy unless significant haemoptysis occurs; if so, surgery is the best therapeutic option or if they are associated with chronic necrotising aspergillosis, antifungal therapy is indicated.¹⁰⁵ Voriconazole is the first line therapy for treatment of invasive pulmonary aspergillosis caused by all common pathogenic species.

Orbital infections

Orbital infections caused by A. flavus are quite uncommon in North America, but frequently seen in South-East Asia, Saudi Arabia and Japan.^106–108 The high prevalence in these areas could be attributed to the fact that Aspergillus spores withstand the hot and dry weather well. Mode of acquisition of the fungus is either endogenous¹⁰⁹ or exogenous and more frequently reported in immunocompromised patients, postcataract surgery or trauma and with i.v. drug use.^{7,9,110–113} The spectrum of infections includes keratitis, scleritis, complicated dacryocystitis and endophthalmitis. Fungal keratitis secondary to A. flavus has been reported in the literature with increasing frequency especially in diabetics, post-trauma and postintraocular surgeries (cataract surgery). Microbiological work up in these cases should include corneal scrapings, corneoscleral biopsy and aqueous humour for fungal stains and culture. Winchester et al. [114] have recently described a new non-invasive way of imaging the cornea in vivo called tandem screening confocal microscopy whereby fungal hyphae were imaged as high-contrast filaments from 60 to 400 microns in length. These authors felt that this could result in rapid diagnosis of fungal keratitis with great implications for management and outcome. Diagnosis of endophthalmitis secondary to A. flavus requires the demonstration of A. flavus in stains, culture and histopathology from vitreous humour or vitrectomy specimens. Thus, the rational approach to the management of orbital infections includes rapid diagnosis sometimes with confocal microscopy or orbital ultrasound,¹¹⁵ fungal stains and culture from tissue or vitreous aspirate, followed by prompt chemotherapy and early surgical intervention, especially if there is disease progression or invasion.¹¹⁰ Systemic itraconazole, intracameral and intravitreous amphotericin B along with systemic amphotericin B have been used with some success.⁸ Newer azoles i.e. voriconazole and posaconazole must be effective, but the data are scant.

Aspergillosis in AIDS

Invasive aspergillosis in AIDS is uncommon and is no longer an AIDS-defining illness. Most cases are isolated case reports and majority of these patients had risk factors such as neutropenia or steroid therapy.^116,117 The role of HIV as an independent risk factor is controversial. Invasive infection occurs in advanced stages of AIDS, particularly with CD4 counts below 100 mm³.^118–121 However, the largest study reported to date on AIDS-associated invasive aspergillosis, not necessarily because of A. flavus, reported the presence of these risk factors only in 48.5% of their patients.¹²² The occurrence of disseminated aspergillosis in 50% of non-neutropenic patients indicates the likelihood of qualitative defects in neutrophils and alveolar macrophages, which are the major cells for host defence against Aspergillus spp.¹²³ Compared with non-HIV patients, tracheobronchial aspergillosis^117,124 and invasion of the central nervous system^125,126 are more likely to occur in the HIV population. As Aspergillus colonisation of the tracheobronchial tree is not uncommon in AIDS,^118,127 biopsy may be the preferred method of diagnosis. Because of the low prevalence, prophylactic therapy is not warranted in these patients.¹²⁸

Cutaneous aspergillosis

Aspergillus flavus has been reported to be the predominant pathogen in both primary cutaneous aspergillosis¹²⁹ where skin is the only organ infected and contiguous cutaneous aspergillosis (e.g. from a paranasal sinus infection).¹³⁰ However, secondary cutaneous aspergillosis, where skin becomes infected as part of a disseminated process, is most often caused by A. fumigatus although cases secondary to A. flavus have been reported as well.^{94,95,104,131} Primary cutaneous aspergillosis from A. flavus, although very rare has been seen in immunocompromised patients and invariably is a catheter-related infection.^4,132 Patients present with erythema, induration and necrotising nodular, nodulo-ulcerative or ulcerative skin or s.c. lesions in the absence of extra-cutaneous or disseminated aspergillosis. Histopathology of skin lesions usually reveals a diffuse and nodular granulomatous mixed infiltrate of lymphocytes, histiocytes, giant cells, numerous eosinophils and neutrophilic microabscesses in the dermis. Septate hyphae can be seen within giant cells and within the dermis. Periodic acid Schiff's stain (PAS) and gomori methanamine silver stain are positive and diagnosis is confirmed by culture. Children with acute leukaemia, who developed primary cutaneous aspergillosis secondary to A. flavus have responded to treatment with topical nystatin and parenteral amphotericin B.¹³³ Prompt intervention with catheter removal, parenteral amphotericin B and oral 5-flurocytosine with debridement has yielded successful outcomes in adult patients. Prognosis, because of the localised disease process, is better than that of the secondary cases. Contiguous cutaneous aspergillosis occurs as an extension of paranasal sinus infection¹³⁴ whereas secondary cutaneous aspergillosis occurs from invasive pulmonary aspergillosis or as part of a disseminated process. In such disseminated infections, treatment with i.v. amphotericin B or itraconazole has yielded poor results. Coexistent indolent paranasal and cutaneous aspergillosis from A. flavus in an immunocompetent patient has been reported.¹³⁵ Cutaneous infection because of A. flavus has also been reported in postoperative wound infection in diabetics.^136,137 Solitary embolic cutaneous aspergillosis, which is an atypical presentation of metastatic aspergillosis, is a distinct entity and so far there have been only four cases, all caused by A. flavus documented in literature.^138,139

All patients were immunocompromised and presented with a single cutaneous lesion that occurred at a distant site in the presence of invasive pulmonary aspergillosis. The lesion progressed from a purpuric papule to a nodule that eventually became necrotic and ulcerated. The lesions most commonly occurred in areas of terminal circulation (head, extremities). Skin biopsy and tissue culture confirmed the diagnosis of A. flavus in all cases. Although multiple skin nodules is the usual presentation of secondary cutaneous aspergillosis in immunocompromised patients, occurrence of a solitary nodule is an atypical presentation, which might partly be attributed to variability in the angioinvasiveness of A. flavus as compared with A. fumigatus and the mycotoxins produced. Secondary cutaneous aspergillosis frequently caused by A. fumigatus also occurs as a result of angiitis, vascular invasion and embolism. However, the lesions are usually multiple and not solitary.¹³⁸

Aspergillus sinusitis

Aspergillus flavus has been reported more frequently than A. fumigatus from patients with fungal sinusitis in Asia and Africa¹⁴⁰ although the incidence in North America has varied among different institutions.⁶ Infection in the immunocompromised host is frequently invasive resulting in ulcerative necrotising lesions of the soft palate, hard palate, nasal cavity and maxillary sinuses and predisposing factors include neutropenia, prolonged antibiotic therapy, presence of nasogastric tube and prolonged hospitalisation.¹⁴¹ In immunocompetent patients, the disease is non-invasive with or without an allergic component. Patients present with pain or pressure over the sinuses, often painless, fever or periorbital swelling. To date, the largest study of invasive fungal sinusitis involving mostly A. flavus, was reported from Nebraska, USA. In this study, the most common presentation was periorbital swelling, sinusitis occurred at a mean of 19 days after admission, most were secondary to disseminated infection.^142,143 A distinct form of fungal sinusitis involves the formation of mycetoma or fungal balls in the sinuses. Patients may be immunocompetent and asymptomatic; diagnosis frequently is made as an incidental finding during imaging for dental procedures. It is more common in the elderly population and maxillary sinus is commonly involved although multiple sinuses could be affected. Both A. flavus and A. fumigatus have been reported as predominant pathogens in a few case reports. Patients have neither an allergic component nor disseminated disease. Surgery is the mainstay of therapy but recurrences are common. Prognosis is good except in cases of sphenoidal sinus involvement. Allergic fungal sinusitis is a unique entity which is a non-invasive inflammatory process of the paranasal sinuses that occurs in immunocompetent people.¹⁴⁴ Non-Aspergillus moulds are frequently the aetiological agents although A. flavus has been isolated in a few cases. Diagnosis is difficult despite the proposed criteria, but critical because unlike invasive infections, antifungal therapy has no role and patients are managed with surgical debridement, systemic and topical corticosteroids. A. flavus has also been reported to cause otomycosis, chronic suppurative otitis media and malignant otitis externa. With the increase in the diagnosis of paranasal infections secondary to A. flavus, Kaplan et al. have recommended the use of standardised orbital ultrasound as a complementary imaging modality to CT scan of sinuses for diagnosing orbital extensions in cases of sinus mycoses. Treatment involves systemic antifungal chemotherapy with or without debridement.

Bone infections

Aspergillus flavus osteomyelitis of long bones and vertebral osteomyelitis have been described both in immunocompetent and immunocompromised patients. Most cases reported in immunocompetent people occurred post-trauma or surgery or in diabetics. Primary bone infections occurred post-trauma whereas secondary osteomyelitis followed either pulmonary aspergillosis or in rare cases endocarditis. Better outcomes were seen in patients who were treated with a combination of surgical debridement and medical therapy^1,145–154 (Table 2).

Table 2.   Infections of the bone and joint caused by Aspergillus flavus

Host	Site of infection	Pathogen	Outcome	References
Closed femur Fracture postop	Chronic osteomyelitis	A. flavus	Surgery + AMB	1
Liver transplant	Coxarthritis Lumbar spodylodiscitis	A. flavus	Liposomal AMB + ITZ + Sx	145
AML	Spinal epidural abscess	A. flavus	AMB + Sx	146
Diabetic	Mycetoma + epidural abscess  + diskitis + osteomyelitis	A. flavus	Decompression laminectomy + ITZ	147
Post-traumatic	R tibial epiphyseal osteomyelitis	A. flavus	AMB	148
No trauma	Acetabular OM	A. flavus	Chemo + Sx	149
Diabetic	Sternal OM	A. flavus	Not available	150
Heart transplant, post-traumatic wound	L tibial epiphyseal OM	A. flavus	AMB + ITZ + Sx	151
Diabetic	Epidural abscess	A. flavus	ITZ + Sx, switched to AMB (died)	152
4 patients-primary pulmonary 4 patients-secondary pulmonary 1 patient-endocarditis	Spondylodiscitis	A. flavus (1) A. fumigatus (8)	ITZ ± 5FC ± AMB No surgery	153
Postcardiac surgery	Sternal wound infection 7 patients-locally invasive 2 patients-fulminant mediastinitis	A. flavus	Chemo + Sx 2 patients with mediastinitis died	154

Miscellaneous infections caused by A. flavus

Several other forms of serious infection reported in the literature include endocarditis, reno-ureteric aspergilloma in a diabetic patient, pulmonary and disseminated infection in a patient with falciparum malaria, primary CNS aspergillosis and aortic graft infections. Most infections were reported in immunocompromised hosts and the mortality rate was high (Table 3).^208–211

Table 3.   Miscellaneous infections caused by Aspergillus flavus

Host	Site of infection	Pathogen	Outcome	Reference
Neurosurgical (abscess, meningitis, hydrocephalus, rhinocerebral)	Disseminated infection	A. flavus	Died	198
Organ transplant (22 patients)	Primary CNS	A. flavus	Poor	199
Diabetic/IVDU	Renoureteric aspergilloma	A. flavus	Sx + ITZ Successful	197,200
Tuberculosis, (38 patients) bronchiectasis	Pulmonary infection	10 patients Aspergillus antibodies; 8/10 A. flavus	Successful	201
AML	Pulmonary + skin	Aflatoxin B1, B2, M	Died	202
Postcardiac surgery	Cardiac	A. flavus	Died	203
Falciparum malaria	Disseminated	A. flavus + Absidia	AMB + VCZ (Failed)	204
ALL	Mitral valve endocarditis, lung abscess	A. flavus	Chemo + Sx (died from embolism)	205
Liver transplant	Cutaneous aspergillosis	A. flavus + CMV	AMB + gancyclovir + topical terbinafine (Died)	206
Aortic stenosis	Aortic graft infection	A. flavus	Died	207

Diagnosis

The diagnosis of A. flavus infections is no different from that of A. fumigatus. The rapidly progressive nature of invasive aspergillosis stresses the need to diagnose and treat these infections at an early stage before dissemination could occur. For suspected invasive pulmonary aspergillosis, CT scan of the thorax would be the best imaging modality as part of the clinical workup and in planning treatment. It may show focal or multiple peripheral pleural lesions, nodules, cavities, consolidation or bilateral diffuse interstitial pneumonitis. Abnormal CT scan should be followed by a bronchoscopy with bronchoalveolar lavage or bronchial washings or by a open lung biopsy. Microscopical examination of tissue from any site using Gomori methanamine silver stain or Calcoflor white stain may reveal the causative organism. However, definitive diagnosis of invasive aspergillosis requires a histopathological section showing tissue invasion of the hyphae along with positive culture from the same site or isolation of fungus from a sterile site (e.g. brain biopsy). In fact, a study from MD Anderson Cancer Institute reported that A. flavus compared with other species of Aspergillus, was significantly more frequently isolated from tissue rather than bronchoalveolar lavage specimens.¹⁵⁵ For pleural-based single lesions needle aspiration of the lesions and for diffuse infiltrates, bronchoscopy with bronchoalveolar lavage are recommended. Some may advocate open lung biopsy with resection, if there is a single nodular lesion. Skin and bone specimens should be sent for fungal stains, culture and histopathology for definitive diagnosis in cases suspected of cutaneous or bone infections respectively. Indium-111-labelled leukocyte scan, a non-specific test, can be used in cases of dissemination when there is difficulty in localising the infection.¹⁵⁶

Various secretory enzymes such as secretory aspartyl proteinase, elastase, lipase and phosphatase as well as cell wall components, namely glucan and galactomannan, have been used as antigens for the detection of Aspergillus infection by enzyme immunoasay in humans. Recent developments in this area of investigation provided some encouraging results although the technique is still suffering from low specificity and sensitivity. The positive aspects of the immunological method are that it is rapid, non-invasive and often requires relatively small amount of sample for the test.^157–162

Major advances in the diagnosis of invasive aspergillosis include the detection of galactomannan antigen by enzyme immunoassay (EIA) in serum (Platelia Aspergillus) and a second non-invasive test, which is the 1,3,-beta d-glucan assay (Glucatell assay kit; Associates of Cape Cod, East Falmouth, MA, USA), both approved by the FDA. The kinetics of both markers in patients with invasive aspergillosis is similar. Both tests have sensitivity, specificity, positive and negative predictive values of 87.5%, 89.6%, 70% and 96.3% respectively. False positive reactions occurred at 10.3% in both tests. A combination of the two tests, when used in the appropriate clinical setting improved the specificity and the positive predictive value of each individual test (to 100%) without affecting the sensitivity and negative predictive values.^163,164 However, biopsy of the involved organ for histopathological examination remains the ideal test.

Adoption of conventional and specialised PCR (e.g. nested PCR, RT-PCR, quantitative RT-PCR) in the recent past has made remarkable contribution to the diagnosis and identification of infectious agents, including those of fungal infections.^165–171 PCR testing is also being introduced in several institutions, some specific for A. flavus, for early diagnosis and empiric therapy in severely ill immunocompromised patients.^172,173 These at present are neither standardised nor commercially available.

Management

With advances in the field of transplantation and novel chemotherapeutic regimens including T-cell depletion, more patients are surviving their cancers but at the expense of an increasing incidence of chronic graft vs. host disease. This in turn has resulted in a dramatic increase in the incidence of life-threatening aspergillosis in the past two decades. More importantly, the morbidity and mortality of these infections despite antifungal therapy remains unacceptably high. The treatment of infections because of A. flavus is no different than the treatment of A. fumigatus. The primary drug in the physician’s armamentarium to treat all invasive forms of aspergillosis is voriconazole. Amphotericin B alone or in combination with azoles or 5-FC has been widely used in the past to treat infections from A. flavus. However, response rates in general have been dismal and clinical failures to amphotericin B (AMB) have been reported. Itraconazole has been shown to be effective in patients who failed AMB treatment.^174,175 Caspofungin, an echinocandin, known to be fungistatic against Aspergillus spp., has also been used with some success in AMB clinical failures.¹⁷⁶ The reason for treatment failure is twofold (i) blood vessel invasion by these fungi causes thrombosis and infarction, thereby rendering it difficult for antifungal drugs to exceed minimal inhibitory concentrations in infected tissues¹⁷⁷ and (ii) immunocompromised host. The newer triazoles including voriconazole and posaconazole have shown excellent fungicidal activity in vitro and in vivo to A. flavus with MICs ranging from <0.002 to 0.5 mg l⁻¹, the lowest reported for posaconazole.¹⁷⁸ Based on quantitative sterol analysis in vitro, using liquid chromatography and mass spectroscopy, posaconazole has been shown to be the most potent inhibitor of 14-alpha-lanosterol demethylase in A. fumigatus and A. flavus when compared with voriconazole or itraconazole including some strains of both species that are resistant to voriconazole or itraconazole.¹⁷⁹

Drug resistance

The effectiveness and the prolonged use of a newly licensed antifungal drug are mainly dependent on the emergence of resistance to this agent among clinical isolates. The emergence of resistance to new drugs in general has great economic significance.^180,181

Resistance can be described as primary (innate) when a fungal pathogen is intrinsically resistant to the antifungal drug, or secondary (acquired) when an organism develops resistance during drug exposure either because of spontaneous mutation or the acquisition of the resistance trait from an external source by genetic transfer. The known cellular and molecular mechanisms responsible for reduced in vitro and in vivo susceptibility to antifungal drugs fall into two broad categories, namely, reduced intracellular accumulation of the antifungal drug compared with that in the susceptible cells, and quantitative or structural alteration of the fungal drug target.

The reduced intracellular drug accumulation occurs either because of efflux of the drug from the cell-mediated by efflux proteins, or because of reduced penetration of the drug into the cell because of selective drug-permeability barrier(s). The efflux proteins belong to two groups, ATP Binding Cassette (ABC) transporters and major facilitators.

A second, less well-known mechanism for the reduced accumulation of antifungal drugs inside the fungal cell is diminished penetration of the drug because of selective permeability barrier(s). Usually, drug target modification dependent mechanism alone or in combination with other resistance mechanism leads to high-level cellular resistance to the antifungal drug.^212–214

As the study of Aspergillus infection caused by A. flavus has become an area of investigation only recently, very little is known about the frequency and the mechanisms of resistance to antifungal drugs in A. flavus at present. Examination of the in vitro susceptibility of A. flavus clinical isolates from the United States, South Africa and India showed that all clinical isolates were highly susceptible to commonly used antifungal drugs such as polyenes, triazoles and echinocandins.¹⁸² However, a comparison of the geometric mean MICs of amphotericin B for A. flavus clinical isolates to those obtained for A. fumigatus showed that the former is consistently at least twofold higher^183–185 perhaps because of innate resistance. This slight increase in MICs of amphotericin B may not have any significant ramifications as the blood level of these antifungal drugs usually exceeds the MICs of these drugs for A. flavus. On the other hand, the clinical isolates of A. flavus are generally highly susceptible to various triazoles.^183,186,187

The research on the molecular mechanisms of antifungal resistance in A. flavus is in its early infancy, consequently very little is known. Only few reports of the clinical isolates of A. flavus resistant to antifungal drugs are available in the literature.^{183–185,188} On the other hand, recently Varanasi et al. [182] and Krishnan et al. [unpublished data, 2005] have selected voriconazole-resistant strains of A. flavus from a drug susceptible clinical isolate in the laboratory. These strains showed higher MIC (MIC ≥16 μg ml⁻¹) to voriconazole. These laboratory-selected isolates also showed higher MICs to other triazoles such as itraconazole, posaconazole and ravuconazole. Seo et al. [188] have isolated an amphotericin B resistant A. flavus isolate in the laboratory. Further studies on the mechanism of resistance to amphotericin B of this isolate showed an altered cell wall chemistry that appears to be responsible for the poor penetration of the drug in the resistant isolate compared with that in the parent strain. These findings suggest that the possible mechanism of resistance to amphotericin B in this isolate is mediated by reduced accumulation of the drug in the cell because of poor penetration.

A. flavus in outbreak setting and prevention

Several studies have been carried out in the past to determine the efficacy of air filtration in hospitals especially in bone marrow transplant units for the prevention of aspergillosis in immunocompromisd patients.¹⁸⁹ Exposure to building construction in the vicinity of the hospital that had no air filtration system was evaluated in another study.¹⁹⁰ Results of these studies indicate that the concentration of Aspergillus spores in filtered air is much less than that in rooms with open windows. Also, investigation of an outbreak of 10 cases of nosocomial invasive infection with A. flavus in an oncology unit in Roswell Park Cancer Institute, Buffalo, New York, demonstrated that the source of outbreak was the high counts of A. flavus conidia in the non-bone marrow transplant wing during the outbreak. Installation of high-efficiency particulate air filtration (HEPA) as an infection control measure in this institution, resulted in a marked reduction of nosocomial aspergillosis in the following 2 years.¹⁹¹ Nosocomial aspergillosis secondary to A. flavus in an outbreak setting has been reported from few other institutions as well.^192–195 Installation of HEPA filters or laminar air flow units has been shown to offer protection to haematopoietic stem cell transplant patients and has proven efficacy in controlling outbreaks because of air contamination with Aspergillus conidia.^191,196

Conclusions

Despite the availability of three different classes of antifungal drugs (polyenes, triazoles and echinocandins), invasive aspergillosis caused by A. flavus continues to be one of the leading causes of morbidity and mortality in immunocompromised patients in several institutions around the world. The recent characterisation and publication of the entire genome of A. flavus, by the United States Department of Agriculture (USDA), marks a new era in research for fungal biology, medical mycology, agricultural ecology and evolution.² It also paves the way for further research into the molecular aspects of the pathogenesis of invasive infections in humans, more research on immune cell interactions^12,64–75 and is also expected to advance the development of therapeutic modalities and devise optimal strategies to control these lethal infections.