Pulmonary aspergillosis: diagnosis and treatment

IPA in haematologic cancer patients

Allogeneic haematopoietic stem cell transplant (HSCT) recipients with graft-versus-host disease (GVHD) and patients with acute leukaemia and prolonged chemotherapy-induced neutropenia (i.e. >10 days) represent the groups of patients at highest risk of IPA with an incidence of 2–8% [14–18]. However, IPA is now diagnosed with increasing frequency among other onco-haematologic populations formerly considered at “low risk”, such as patients with lymphoma, multiple myeloma or chronic leukaemia [8, 19–21]. Moreover, novel anti-cancer therapies, such as the Bruton tyrosine kinase inhibitors (e.g. ibrutinib) have been associated with an increased risk of IPA [22].

Neutropenic haematologic cancer patients usually have pathognomonic radiological findings on chest CT, such as well-circumscribed dense lesions (nodules or masses) with or without a surrounding area of ground-glass opacity due to haemorrhage known as the halo sign [10]. The halo sign usually represents an early stage of the disease, while the air crescent sign (partial cavitation) and the presence of a cavity are later-stage radiological manifestations with poorer prognosis [10, 23].

In haematologic cancer patients, the yield of direct exam and cultures for Aspergillus detection is notoriously low (around 20% and 50%, respectively) [24–26]. Nonculture methods, in particular GM or PCR in serum or BAL, are the cornerstone of IPA diagnosis and are in more than 50% of cases the only mycological markers of IPA [8, 12, 25, 27]. Their sensitivity and specificity are presented in table 2. Overall, GM and PCR exhibit comparable diagnostic performances and their combined use in serum or BAL samples results in the highest sensitivity and specificity [28–30]. BDG testing in serum is not specific for Aspergillus species and is associated with some false-positive results in this population, notably among patients who have received intravenous immunoglobulins [11, 31, 32].

The epidemiology and diagnostic approach of IPA has changed with the increasing use of posaconazole prophylaxis in high-risk patients [33, 34]. Therefore, serial screening of GM in serum is not recommended in patients under posaconazole prophylaxis because of the low incidence of IPA, which results in low positive predictive value of the test [34, 35]. However, a pre-emptive approach with serial serum GM screening (e.g. twice per week) represents an alternative to systemic antifungal (AF) prophylaxis in these high-risk patients [34]. Chest CT scan should be performed in case of persistent or recurrent febrile neutropenia despite broad-spectrum antibiotics or in case of a positive serum GM. Bronchoscopy with BAL should be considered in the presence of a lung lesion consistent with IPA at CT scan. Indeed, BAL samplings are associated with the highest sensitivity and specificity to detect IPA (table 2) [12]. An algorithm for the diagnostic work-up of IPA in high-risk haematologic cancer patients is presented in figure 2. Patients with a positive serum GM have a worse prognosis, because of more advanced disease, which suggests that clinicians should have low threshold to perform chest CT for early detection of IPA [24].

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Diagnostic approach of invasive pulmonary aspergillosis (IPA) in high-risk haematologic cancer patients. AF: antifungal; BAL: bronchoalveolar lavage; CT: computed tomography; EORTC/MSGERC: European Organization for Research and Treatment of Cancer and Mycoses Study Group Education and Research Consortium; GM: galactomannan; GVHD: graft versus host disease; HSCT: haematopoietic stem cell transplantation; IA: invasive aspergillosis; MRI: magnetic resonance imaging; ODI: optical density index; PCR: specific polymerase chain reaction for Aspergillus species.

For other haematologic cancer patients at lower risk of IPA (e.g. chronic leukaemia, lymphoma, autologous HSCT recipients), anti-mould prophylaxis or serial GM screening are not recommended, but the threshold of clinical suspicion should be low to trigger chest CT and further investigations as IPA diagnosis is often missed in this population [19].

IPA in solid organ transplant recipients

The incidence of IPA among solid organ transplant (SOT) recipients is estimated to be 1–3% [36, 37]. It is higher among heart and lung than in other organ transplant recipients. The diagnosis of IPA in this population is difficult because of the lower specificity of radiological findings and lower sensitivity of mycological tests compared to those observed in haematologic cancer patients. Chest CT findings of IPA in SOT are not limited to well-circumscribed lung lesions (e.g. nodules or masses), but include a wide range of nonspecific abnormalities that cannot be distinguished from bacterial or viral pneumonias, such as tree-in-bud patterns, lobar infiltrates or ground-glass opacities [10].

The sensitivity of GM in serum is lower (around 40%) in SOT than in haematologic cancer patients [38]. BDG has poor sensitivity and specificity in this population [31, 39]. However, overall performance of GM and PCR in BAL is comparable to that reported in haematologic cancer patients with notably a similar high performance for their combined use [40, 41].

There is no recommendation for serial monitoring of serum fungal biomarkers in SOT patients because of their limited sensitivity and specificity [34]. In case of clinical suspicion (e.g. suggestive chest CT lesion), bronchoscopy should be performed and combined use of mycological tests (culture, GM, PCR) in BAL is the cornerstone for IPA diagnosis.

IPA in other immunocompromised hosts

Data about the performance of culture and nonculture diagnostic tests of IPA among patients with other immunosuppressive conditions (e.g. auto-immune diseases, corticosteroid therapy, treatment with Bruton tyrosine kinase inhibitors) are lacking. However, it is expected to be comparable to that observed in other nonneutropenic populations, such as SOT recipients.

IPA in ICU patients

The incidence of IPA in ICU is highly variable. Patients with severe influenza or COVID-19 are at the highest risk with reported incidences ranging from 5 to 30% [42]. There is more evidence supporting the association of IPA and influenza with one control study demonstrating a significantly higher incidence of IPA among ICU patients with severe influenza compared to those with other community-acquired pneumonia [6]. Moreover, evidence of angio-invasion, based on a positive GM in serum or histopathology, is more frequent in influenza than in COVID-19 [42–44]. In the latter, the distinction between Aspergillus colonisation and true IPA is difficult [45, 46]. However, recovery of Aspergillus in a respiratory sample by direct or indirect diagnostic tests seems to be a marker of poor prognosis in both influenza and COVID-19 patients [4, 6]. Other conditions that may predispose to IPA among ICU patients include liver cirrhosis, extracorporeal membrane oxygenation and short courses of corticosteroids [5, 47, 48].

IPA should be suspected in intubated patients with deteriorating respiratory conditions despite broad-spectrum antibiotic therapy. Chest CT findings are usually not distinguishable from viral or bacterial pneumonias [5, 42, 49, 50].

The EORTC-MSGERC criteria for IPA do not apply in ICU patients who do not have host criteria of immunosuppression in most cases. Customised definitions have been proposed for IPA in ICU, and specifically for IPA in patients with influenza or COVID-19 [49–51]. While the distinction between Aspergillus colonisation of the airways and true IPA is difficult to assess in this setting, the diagnostic criteria usually rely on a positive test for Aspergillus (direct exam, culture, GM or PCR) in a BAL sample. The significance of positive results in nonbronchoscopic respiratory samples (e.g. bronchial aspirates) is more difficult to interpret. GM in serum has a sensitivity around 50–60% for IPA detection in influenza [6, 44], but of less than 15% in COVID-19 [42].

Triazoles

Mould-active triazoles include itraconazole, voriconazole, posaconazole and isavuconazole. These AF agents inhibit the biosynthesis of ergosterol and are fungicidal against most Aspergillus species with some rare exceptions such as cryptic species of section Fumigati (e.g. A. lentulus, A. udagawae) and all species of section Usti (e.g. A. calidoustus), which account for a very low proportion of IPA (<3%) [59, 60]. Triazoles have the advantage to be available as intravenous and oral formulations. Their main side-effect is hepatotoxicity, which is more frequent in patients treated with voriconazole and posaconazole (10–15%) than in those treated with isavuconazole (<10%) [25, 61]. Hepatotoxicity may require treatment interruption but is usually reversible. Triazoles are substrates and inhibitors of cytochrome P450 isoenzymes, which results in multiple drug–drug interactions [62]. These interactions are a particular concern for voriconazole and may require dosage adaptations of other drugs, but they rarely prevent the use of triazoles. With the exception of isavuconazole, triazoles may prolong the QT interval and require electrocardiographic monitoring. Significant QT prolongation has been observed in about 15% of onco–haematological patients and is associated with other medications affecting the QT interval and/or electrolytic disorders in most cases [63]. Life-threatening complications (torsades de pointes) are very rare events. Therapeutic drug monitoring (TDM) is recommended for itraconazole, voriconazole and posaconazole because of inter- and intra-individual variability of serum concentrations [34]. For voriconazole, trough concentrations of 1 to 5.5 mg·L⁻¹ have been proposed based on pharmacodynamic studies correlating low and high concentrations with failure of therapy and toxicity (visual or neurological disturbances), respectively [34, 64, 65]. Genetic polymorphisms in the CYP2C19, the cytochrome P450 isoenzyme involved in the primary metabolism of voriconazole, can explain inter-individual variations between rapid and poor metabolisers [66]. For posaconazole, important variations of serum concentrations have been observed with the oral suspension formulation and have been related to high gastric pH (e.g. co-administration of proton pump inhibitors, type of food) [67]. More stable pharmacokinetic profiles are observed with the new oral formulation (tablets) and the intravenous formulation [68]. However, TDM is still recommended with targeted trough concentrations of 0.7 mg·L⁻¹ and 1.0 mg·L⁻¹ for prophylaxis and treatment, respectively [34]. For isavuconazole, TDM is not routinely recommended because of more stable pharmacokinetic profiles and lack of correlation between trough concentrations and outcomes [34, 69].

Triazoles represents the mainstay for the treatment of aspergillosis [2, 33, 34]. While voriconazole is recommended as the first-line therapy of IPA [33, 34], new azoles (posaconazole, isavuconazole) have demonstrated similar efficacy in randomised clinical trials [25, 61]. Isavuconazole is of particular interest as it is associated with less hepatotoxicity, fewer drug interactions and more stable pharmacokinetic profiles (i.e. no need for TDM) than voriconazole [25, 70]. The global emergence of azole resistance in A. fumigatus, thought to be associated with the widespread use of fungicides in the environment, is of particular concern [71]. Various mutations have been described in the azole target gene (cyp51A), the most frequent ones being the TR₃₄/L98H and the TR₄₆/Y121F/T289A mutations, which are associated with pan-azole resistance [72]. Patients receiving first-line voriconazole therapy for IPA due to an azole-resistant A. fumigatus have significantly higher mortality rates compared to those infected with an azole-susceptible isolate [73].

Amphotericin B

This AF agent binds to ergosterol, which results in altered integrity of the cell membrane with loss of protons and monovalent cations [74]. The AF activity of amphotericin B may also result from production of reactive oxygen species and oxidative damage [75]. It has a broad spectrum of AF activity and is fungicidal against most Aspergillus species with the exception of A. terreus [74, 76]. Acquired resistance among other Aspergillus species, such as A. fumigatus, is rare and mechanisms are poorly elucidated [77].

Over the last two decades, conventional amphotericin B deoxycholate has been substituted by safer lipid formulations of amphotericin B, such as liposomal amphotericin B, amphotericin B lipid complex or amphotericin B colloidal dispersion.

The use of amphotericin B is limited by its adverse events and its availability as an intravenous formulation only. Nephrotoxicity is the main adverse effect, which was observed in one third of patients treated with amphotericin B deoxycholate [78]. However, novel amphotericin B lipid formulations are associated with reduced nephrotoxicity (10–15%) [78]. Other side effects, such as infusion-related reactions (fever, rigors) and hypokalaemia are also significantly less frequent with the lipid formulations, notably liposomal amphotericin B [78, 79].

One randomised study showed that amphotericin B deoxycholate was associated with lower rate of therapeutic success and more drug-related adverse events compared to voriconazole [80]. Although data from randomised controlled trials are lacking for liposomal amphotericin B, this drug has been associated with success rates similar to those observed with voriconazole for IPA treatment [81].

Liposomal amphotericin B is a second-line therapy of IPA and the first-line therapy for azole-resistant IPA [34]. It is also recommended as initial empirical therapy for suspected or possible invasive mould infection because of its broader spectrum of activity against pathogenic moulds including the Mucorales and azole-resistant Aspergillus species [34]. The local epidemiology regarding the prevalence of these more resistant pathogens should be taken into consideration in the choice of initial empirical AF therapy.

Echinocandins

Echinocandins inhibit the beta-glucan synthase, which results in loss of the fungal cell wall integrity [82]. Currently licensed echinocandins are caspofungin, anidulafungin and micafungin, which are available only as intravenous formulations. Rezafungin is a novel molecule of this class with prolonged half-life that can be administered once weekly [83]. Echinocandins have limited in vitro activity against Aspergillus species with a fungistatic effect resulting in blunted hyphae and growth inhibition [82]. In vitro and in animal models, caspofungin displays a paradoxical effect (i.e. decreased efficacy at increased concentrations), of uncertain clinical relevance [82]. Acquired echinocandin resistance may occur in A. fumigatus via mutations in the beta-glucan synthase gene but this has been very rarely observed in clinical isolates [82, 84].

In clinical practice, variable success rates have been reported with echinocandin monotherapy and there is no randomised trials comparing the efficacy and safety of echinocandins versus amphotericin B or triazoles [82, 85]. Currently, echinocandins are not routinely recommended for IPA treatment, except in particular situations when underlying conditions or toxicity prevents the use of triazoles or amphotericin B formulations [33, 34]. Echinocandins may be used in combination therapy with voriconazole in severe IPA [34]. While this combination demonstrated some synergism in vitro and in animal models [82, 86, 87], its superiority over voriconazole monotherapy could not be clearly demonstrated in a randomised trial showing only a trend towards better survival with the combination and a significant effect limited to the subgroup of patients with a positive GM [26]. This combination also represents a therapeutic option for the treatment of azole-resistant IPA (as an alternative to liposomal amphotericin B), but clinical efficacy in this setting is not demonstrated [88]. Some pre-clinical (in vitro and murine model) and clinical data (single pilot study) also suggest a synergistic interaction between echinocandins and amphotericin B [82, 89].

Novel AF agents

Because of the limited therapeutic options against azole-resistant IPA, novel AF drugs with potent and broad anti-Aspergillus activity are warranted. Olorofim is an inhibitor of the fungal dihydroorotate dehydrogenase, an enzyme involved in pyrimidine biosynthesis, which has a potent activity with a time-dependent fungicidal effect against all Aspergillus species including those with intrinsic or acquired azole resistance [58]. The prodrug fosmanogepix, which is converted to its active moiety manogepix by systemic phosphatases, is an inhibitor of an enzyme in the glycosylphosphatidylinositol anchor biosynthetic pathway (Gwt1) [58]. It has potent fungistatic activity against all Aspergillus species including those with intrinsic or acquired azole resistance. Both compounds have the advantage to be available as intravenous and oral formulations. They are currently in phase II clinical trials for the treatment of IPA [58]. Ibrexafungerp is a beta-glucan synthase inhibitor with a structure distinct from echinocandins that can be administered by oral route [58]. Its spectrum of activity against Aspergillus species is comparable to that of echinocandins [58]. A phase II trial comparing ibrexafungerp given in combination with voriconazole versus voriconazole monotherapy for the treatment of IPA is ongoing [58].

Within existing drug classes, other drugs with improved pharmacologic properties (e.g. tetrazoles, encochleated amphotericin B) are in clinical phases of investigation [58].

IPA

AF therapy should be initiated promptly when IPA is suspected. When a bronchoscopy is planned, AF therapy should be started after the procedure whenever possible, but should not be delayed by more than 24 h. The therapeutic approach of IPA in the different subsets of immunocompromised patients (e.g. haematologic cancer patients, SOT, ICU patients) does not substantially differ. Triazoles (voriconazole, isavuconazole or posaconazole) should be privileged for proven or probable IPA and liposomal amphotericin B for possible IPA to ensure broad coverage including the Mucorales and azole-resistant Aspergillus species. Other parameters, such as the baseline liver and kidney functions and the co-medications with potential drug–drug interactions may also influence the choice of initial AF therapy as described above. Combination therapies of triazoles and echinocandins should be limited to patients with severe or refractory IPA. Surgical interventions are rarely necessary for IPA unless there is a need for source control because of involvement of adjacent structures (e.g. pleura, mediastinum), extrapulmonary foci (e.g. cerebral aspergillosis), life-threatening haemoptysis or nonresponsive disease. Bone-marrow recovery or reduction of immunosuppression are important outcome determinants. Tapering of immunosuppressive therapies should be considered very carefully in a multidisciplinary approach to avoid graft rejection, GVHD or worsening of underlying immune diseases.

The duration of therapy is not well defined, but it should be at least 6–12 weeks depending on the recovery of immunosuppression and the evolution of CT lesions that should be carefully monitored [33, 34]. Some case series have addressed the role of positron emission tomography/CT in monitoring response to therapy of invasive fungal infections with interesting results, which should be confirmed in larger datasets [10, 90]. Secondary prophylaxis should be considered in patients with prolonged immunosuppression [34]. Shorter treatment duration (e.g. 4–6 weeks) may be considered in immunocompetent ICU patients (e.g. influenza and COVID-19) once they are discharged from ICU [46].

Chronic pulmonary aspergillosis (CPA)

The management of CPA is complex and requires a multidisciplinary approach involving pulmonologists, thoracic surgeons, radiologists and infectious diseases specialists. As CPA may present with very heterogeneous radiological patterns and degree of extension, indications for surgery should be carefully assessed on an individual basis considering the number and size of the lesions and the patient's underlying structural lung disease and/or degree of lung function impairment [2]. In case of surgery, pre-operative AF therapy may be warranted for complex surgical procedures to prevent spillage of the fungus (e.g. within pleural cavity) [2]. Post-operative AF therapy should be considered in case of partial resection of the lesion(s), complex procedures with risk of spillage or when histopathology or cultures show presence of fungal elements invading the adjacent lung parenchyma [2].

In the absence of surgery, AF therapy may prevent progression of the lesions (e.g. fibrosis) and life-threatening complications, such as haemoptysis, and improve respiratory conditions, quality of life and outcomes [2, 91]. Long-term suppressive AF therapy is usually required as therapeutic response is very slow and only partial in most cases. Regular chest CT (every 3 or 6 months) is indicated to assess therapeutic response [2]. Oral triazoles are the cornerstone of CPA treatment. Itraconazole may be used as first-line therapy because of its low cost and good safety profile [2, 91]. Voriconazole (or posaconazole) can be administered for more severe disease [2, 92, 93]. Isavuconazole can be used in case of toxicity or drug–drug interactions. Patients with long-term triazole therapy should have regular dermatologic control (every 6 months) for the early detection of skin cancer [94]. Echinocandin therapy may be considered in case of intolerance or resistance to triazoles [2, 95]. Rezafungin (administered only once a week) or the novel oral beta-glucan synthase inhibitor ibrexafungerp may represent interesting alternative treatments although data supporting their efficacy in CPA are currently lacking.

For patients who are not eligible for surgery and experience failure of AF therapy with recurrent haemoptysis, direct instillation of amphotericin B or a triazole (itraconazole, miconazole) in the aspergilloma cavity has shown some efficacy [2, 96].

Haemoptysis is the most feared complication and may require administration of tranexamic acid or arterial embolisation in the most severe cases [2]. Surgery should be considered in case of severe or recurrent haemoptysis [2].

Because of the paucity of clinical trials, management of CPA mainly relies on experts’ opinions [2]. A multicentre collaboration has been recently initiated to establish an international registry (CPAnet), which is expected to improve our understanding of the disease and to standardise therapeutic approaches [97].