Fungal epiglottitis is a very rare condition and is usually attributed to either Aspergillus or Candida spp. Only three cases of Aspergillus epiglottitis have been reported in the literature to date, two in the immunocompromised population, HIV and haematological malignancy, (Bolivar et al 1983 and Sriskandabalan and Roy 1996), and one in the immunocompetent (Durell et al 2011). In the immunocompromised population necrosis of the epiglottis with ulceration was seen, with complete destruction of the epiglottis and replacement by white necrotic tissue with gross laryngeal oedema being described by Bolivar et al (1983). The presence of a pseudomembrane in the absence of ulceration has also been described (Hu et al 2012). Both Aspergillus flavus (Bolivar et al 1983) and Aspergillus fumigatus (Durrel et al 2011, Sriskandabalan and Roy 1996) have been isolated as causal pathogens and co-infection with bacterial pathogens, for example Klebsiella pneumoniae, has also been reported (Bolivar et al 1983). It is possible that co-existent bacterial infection facilitates fungal invasion through superficial tissue destruction with altered oropharyngeal flora following broad spectrum antibiotic use, as is common in those with profound immunosuppression, also contributing to an increased risk of invasive disease.
As epiglottitis can cause upper airway obstruction mortality is high in the reported literature. However securing a definitive airway early with the introduction of systemic antifungal therapy can prove life saving. Systemic amphotericin therapy has been used, with variable success, and it is probable that both azole and echinocandin therapy may play a role although there are no specific data in cases of Aspergillus epiglottitis at present to support their use in the acute phase of infection. Voriconazole is likely to be superior to amphotericin B, as it is for pulmonary and cerebral aspergillosis. Azole therapy has, however, been described as consolidation therapy following systemic treatment in the HIV population (Sriskdabalan and Roy 1996).
Aspergillus infection involving the larynx is rare and is associated with two distinct patterns of disease in two distinct patient groups, the immunocompromised and the ostensibly immunocompetent. Histologically patterns differ with laryngeal biopsies in the immunocompetent demonstrating no vascular invasion with fungal hyphae seen throughout the specimen. In contrast in invasive disease abundant hyphae, vascular invasion, micro-thrombosis with tissue infarction, necrosis, oedema and haemorrhage are seen (Benson-Mitchell 1994); this pattern most commonly being seen in the immunosuppressed. These histological patterns lead to two distinct clinical presentations as discussed below although in reality there is probably a spectrum of histological and clinical presentation according to individual local factors and the degree of impairment of host response.
Aspergillus laryngitis in the immunocompromised
Cases described in the immunocompromised population tend to centre on those with haematological conditions, including haematological malignancy, and HIV (Young et al 1970, Kingdom et al 1996, Delbrouck et al 1998, Barnes et al 2001, Nagasawa et al 2002, Athanassiadou et al 2005, Williams et al 2013). Common presenting features in this group include fever, cough, stridor and dysphonia. Respiratory distress becomes a predominant feature as the disease progresses with the need for a definitive airway, in the form of either endotracheal intubation or tracheostomy, in all reported cases. Critical airway narrowing can often be identified by direct laryngoscopy but can also be seen radiologically with appropriate cross sectional imaging (Williams et al 2013).
Aspergillus fumigatus is the most commonly reported pathogen (Barnes et al 2001, Nagasawa et al 2002, Athanassiadou et al 2005) and systemic intravenous therapy, in the form of either amphotericin B or voriconazole has been used with some success. Nebulised amphotericin has also been reported as an adjunctive therapy in this patient group (Delbrouck et al 1998, Barnes et al 2001). High dose corticosteroid therapy has also been used with poor clinical outcomes despite several patients initially improving following administration. This response probably reflects either attenuation of an allergic upper airways response or, more likely, reduction in laryngeal oedema with an associated increased risk of progression to invasive disease, at least in part contributing to the subsequent deterioration seen in all cases. (Nagasawa et al 2002). Whilst dexamethasone remains critical in the early management of stridor of unknown aetiology it’s use in known or suspected Aspergillus upper airway infection should be avoided if at all possible. Radical surgery, in the form of total laryngectomy and hemithyroidectomy, has also been described in the case of an A. flavus subglottic abscess failing to respond to broad spectrum antimicrobial cover and percutaneous drainage (Lapointe et al 2004).
Despite treatment, mortality in this group remains high either from laryngeal/tracheal obstruction (Nagasawa et al 2002), detachment of a necrotic laryngeal pseudo-membrane formed as a consequence of infection (Williams et al 2013) or disseminated disease, often pulmonary with associated respiratory failure (Delbrouck et al 1998, Nagasawa et al 2002).
Aspergillus laryngitis in the non-immunocompromised
Laryngitis in this patient group remains rare but is more commonly described than in those with severe immunocompromise, again with A. fumigatus (Fairfax at al 1999) and A. flavus (Morelli et al 2000) being isolated commonly. Subtle defects in immune function may play a role in increasing susceptibility but it is also likely that local factors, directly impacting on laryngeal structures, play a more important role in disease development (Ogawa et al 2002). Local factors potentially implicated in disease development include mucosal damage secondary to oral sex (Ran et al 2012), previous radiotherapy or laser therapy to the vocal cords/larynx (Ogawa et al 2002), vocal cord trauma due to repetitive over use/vocal cord cysts (Liu et al 2010), gastro-pharyngeal reflux (Ravikumar et al 2014), smoking (Ravikumar et al 2014) and high dose inhaled corticosteroid therapy (Fairfax 1999, Liu et al 2010, Ravikumar et al 2014). Reported systemic factors causing subtle immune changes with resulting localised laryngeal disease include recent pregnancy (Villanueva et al 2013), diabetes (Ogawa et al 2002, Liu et al 2010), CD4 T cell lymphopenia with reduced CD4:CD8 ratio (Nakahira et al 2002), long term oral corticosteroid therapy (Morelli et al 2000, Ogawa et al 2002) and Felty’s Syndrome (Morelli et al 2000). An association with extensive laryngeal papillomatosis has also been described which combines local factors and subtle immunodeficiency (Dean 2001). In reality it is likely that many of these factors co-exist and no studies to date have fully investigated individuals with pre-existing “normal” immune function for subtle immune defects.
In contrast to the immunocompromised group the majority of these patients have a very indolent course with prolonged dysphonia, progressing to complete aphonia in some cases (Nong et al 1997,Fairfax 1999), vocal fatigue and cough. Fever is a less common finding in this group but has been reported. In keeping with the variable histological appearances produced by laryngeal Aspergillus infection (granulomata formation, diffuse necrosis and, most commonly, pseudomembrane formation (Rao 1969)), the morphological appearances of Aspergillus spp. disease at laryngoscopy vary. Ulceration, which may be haemorrhagic, or mass lesion alone or ulceration in association with a vocal cord mass (Rao 1969, Young et al 1970, Ferlito et al 1974, Kheir et al 1983, Benson-Mitchell 1994, Nong et al 1997, Fairfax 1999, Villaneuva et al 2013) and white plaques involving the vocal cords and associated laryngeal structures (Nong et al 1997, Morelli et al 2000, Nakahira et al 2002, Liu et al 2010, Ran et al 2012, Ravikumar et al 2104) have all been commonly described. More rarely black speckling in association with the presence of a white eschar (Dean et al 2011) and white spherical submucosal cysts have also been described; the latter also being described as a sub-mucosal aspergilloma (Wittkopf et al 2006, Liu et al 2010). Oedema, hyperkeratosis and erythema of the surface epithelium have also been described as associated, although none are specific, morphological features (Delbrouck et al 1998). Laryngeal cartilage may also be exposed in the presence of infection as a result of diffuse necrosis, this being reported in three out of eight patients in a small series by Ogawa et al (2002). Purulent chondritis of the laryngeal cartilages has also been described secondary to A. fumigatus with demonstration of abscess formation between perichondria on CT imaging (Eliashar et al 2005).
Given the symptoms and morphological appearances of laryngeal aspergillosis malignancy is often considered as a primary diagnosis (Kheir et al 1983, Benson-Mitchell 1994, Beust et al 1998 and Ogawa et al 2002). Histology may be difficult to interpret with acanthosis and pseudoepitheliomatous hyperplasia mimicking malignancy and, conversely, the combination of fungal hyphae and vascular emboli, with subsequent necrosis, may mask malignant disease in the setting of recurrent laryngeal cancer (Ogawa et al 2002)
Despite the sometimes extensive nature of lesions present, vocal cord and laryngeal mobility often remains intact with Ogawa et al (2002) describing preserved mobility in 50% of patients. Kheir (1983), Liu et al (2010) and Ravikumar et al (2014) also report normal vocal cord movement despite the presence of vocal cord lesions in isolated cases.
Treatment is not mandatory in this patient group and there are reports of complete recovery with no intervention following diagnostic biopsy (Rao 1969, Benson-Mitchell 1994). Similarly resection of the lesion alone may prove curative (Kheir et al 1983, Liu et al 2010, Villaneuva et al 2013). Resection has also been achieved via cautery of the affected area with CO2 laser therapy (Ogawa et al 2002).
Medical management of Aspergillus laryngitis focuses on topical and systemic therapy. Topical therapy, in the form of 10mg amphotericin lozenges four times daily, has been reported to be successful in isolation (Fairfax 1999) with gargled amphotericin proving a useful adjunct following surgery (Ogawa et al 2002). Systemic therapy with both amphotericin (Morelli et al 2000) and azole therapy (Dean 2001, Ogawa et al 2002, Nakashira et al 2002, Liu et al 2010, and Ran et al 2012) is described, with azole therapy used as maintenance therapy following iv amphotericin in one case (Morelli et al 2000). The duration of optimal treatment is not clear although oral azole has been used for between 21 and 28 days with good results.
None of the cases described relapsed following treatment and, although optimal treatment strategies are not defined, it is likely that therapy will be delivered on an individual patient basis tailored to the clinical history and degree of underlying immune suppression.
Epiglottitis and laryngitis caused by Aspergillus are rare manifestations of upper airways disease affecting patients with and without immunosuppression. In patients with immunosuppression or rapidly progressing disease then systemic therapy with voriconazole should be instituted, unless contraindications. There are no data on echinocandin use in Aspergillus upper airway infection. However in patients with Aspergillus laryngitis and no known, or mild, immune suppression excision biopsy or cautery of the affected area alone may prove curative.
University Hospital of South Manchester and University of Manchester