The risk for invasive pulmonary aspergillosis (IPA) is increased in immunocompromised patients. There are also increasing reports of immunocompetent patients with respiratory viral and IPA co-infection, particularly in patients admitted to intensive care units (ICU) with severe influenza1 and more recently in patients with coronavirus disease 2019 (COVID-19).2-7 We report the emergence, clinical characteristics and outcomes of patients with COVID-19-associated pulmonary aspergillosis (CAPA), together with a recommended approach for screening and management in the Australian setting.

Between March and August 2020, 50 patients with laboratory-proven severe adult respiratory syndrome coronavirus-2 (SARS-CoV-2) infection required ICU admission at our facility. Of these, four (8%) patients had concurrent possible IPA (Table 1). All four patients required mechanical ventilation and were administered intravenous dexamethasone; one patient also received intravenous remdesivir. Chest computed tomography performed in three patients revealed ground-glass opacity and bilateral consolidation without nodules, halo sign or cavitation (Fig. 1). After a median of 5 days in ICU (range 2–9 days), Aspergillus spp. were isolated in tracheal aspirates of three patients (Aspergillus fumigatus, n = 2; Aspergillus terreus, n = 1), while Patient 4 had two tracheal samples with positive Aspergillus polymerase chain reaction (PCR) and negative fungal culture. Voriconazole treatment was commenced in the setting of declining respiratory function in three patients. Patient 2 was palliated before the initiation of antifungal treatment.

Table 1. Clinical characteristics of ICU patients with COVID-19 and pulmonary aspergillosis Patient 1 Patient 2 Patient 3 Patient 4 Comorbidities Asthma, poorly controlled type 2 diabetes mellitus, hypertension Child-Pugh C liver cirrhosis from chronic hepatitis B End-stage renal failure from type 2 diabetes, renal transplant, myeloma, hypertension, obstructive sleep apnoea Hypertension, renal cell carcinoma (previous nephrectomy), ischaemic heart disease, ex-smoker Time between COVID-19 symptom onset and presentation to hospital 6 days 9 days Already in hospital for 4 days for other reason before symptom onset 5 days CT chest changes Widespread bilateral ground-glass opacity and consolidation at bilateral lower lobes. No definite evidence of pulmonary nodules or cavitation Patchy ground-glass opacities and smooth interlobular septal thickening within both lungs with upper zone predominance. No halo signs or cavitation Extensive ground-glass changes throughout both lungs with bibasal consolidation. No evidence of nodules, halo signs or cavitation Not done Species isolated from tracheal aspirates (culture) Aspergillus fumigatus complex Aspergillus fumigatus complex Aspergillus terreus complex Culture negative Time between ICU admission and positive Aspergillus test 2 days 9 days 6 days 3 days COVID-19-specific therapy Intravenous dexamethasone 6 mg/day (7 days) Intravenous dexamethasone 6 mg/day (7 days) Intravenous dexamethasone 6 mg/day (10 days) Dexamethasone (oral) 6 mg/day (10 days) and remdesivir (6 days) Time between dexamethasone and positive Aspergillus test 2 days 9 days 8 days 6 days Serum GM/Aspergillus PCR (subsequent tracheal aspirates) Done only after the treatment is initiated: negative for both GM and Aspergillus PCR Not done Done only after the treatment is initiated: negative for both GM and Aspergillus PCR Done prior to treatment: 2× tracheal aspirate samples positive for Aspergillus PCR; 2× serum GM negative Treatment Voriconazole (23 days) No Voriconazole (13 days) Voriconazole (9 days) Response to treatment Improvement in pulmonary infiltrates on repeat CT chest — Unclear Unclear Outcome Alive Died Died Died COVID-19, Coronavirus disease 2019; CT, computed tomography; GM, galactomannan; ICU, intensive care unit; PCR, polymerase chain reaction.

Computed tomography images of the chest of Patients 1, 2 and 3, showing ground-glass opacities in both lung fields, with no evidence of nodules, halo signs or cavitation.

All-cause mortality among COVID-19 patients admitted to ICU was 22% (11/50 patients). In contrast, all-cause mortality occurred in 3 of 4 patients in those with CAPA compared with 8 of 46 patients in those without aspergillus co-infection.

To estimate the relative disease burden of IPA in ICU patients during the COVID-19 pandemic, a baseline period (January 2017–February 2020) was compared with the pandemic period (March–August 2020). In the pre-COVID-19 period, 13 cases of IPA were identified in 9277 ICU-admitted patients, corresponding to an IPA rate of 1.4/1000 admitted patients (46% of cases had respiratory virus co-infection). In contrast, 7 cases of IPA were identified in 1171 ICU-admitted patients during the pandemic period, corresponding to a rate of 6.0/1000 admitted patients (71% with respiratory virus co-infection, and the majority (4/5) of these associated with SARS-CoV-2 infection).

Internationally, variable but significant rates (3.2–35%) of CAPA have been reported in patients with acute respiratory distress syndrome (ARDS) complicating SARS-CoV-2 infection, with mortality as high as 65%.2-7 We observed an 8% rate of CAPA in critically ill COVID-19 patients with ARDS admitted to our ICU, with mortality in three of four patients. Median time between ICU admission and CAPA diagnosis has been reported as 6 days,2 consistent with our observation of early onset infection.

Damage to the lung epithelium by SARS-CoV-2 and defective fungal host responses secondary to the hyperinflammatory response to the virus are thought to predispose to Aspergillus infection.2 Other proposed risk factors include corticosteroid use in patients with ARDS, broad-spectrum antibiotics and underlying structural lung disease.2 In particular, there has been increased use of dexamethasone since publication of the RECOVERY trial.8 Concomitant use of corticosteroids with interleukin-6 receptor antagonist, such as tocilizumab, may also increase patients' susceptibility to pulmonary Aspergillus infection.9, 10

It can be difficult to differentiate Aspergillus colonisation from infection in non-neutropenic ICU patients who do not have underlying immunocompromised states and in whom typical radiological findings of invasive fungal disease may be absent. This highlights the need for an agreed consensus definition for CAPA, which was recently proposed, relying on entry criteria of ICU admission for respiratory distress with a positive COVID-19 test temporally related to ICU admission. Proven, probable or possible CAPA have been proposed based on sample validity and diagnostic certainty.11 Despite this, there is ongoing debate if CAPA represents an invasive disease post SARS-CoV-2 infection as seen in influenza-associated pulmonary aspergillosis, or whether it represents only colonisation of patients' airways.12, 13 Understandably, it has been challenging to prove the association between COVID-19 and IPA, even in the setting of post-mortem evaluation;14 hence, the majority of the reported cases to date were categorised as probable or putative CAPA.13

The authors of the consensus regarded bronchoscopy as the cornerstone of CAPA diagnosis; however, due to resource constraint in ICU, coupled with the need to reduce transmission risk related to aerosolisation, bronchoscopy is not routinely performed in many institutions including ours. Jabeen et al. criticised this emphasis on bronchoscopy samples as such strict criteria may lead to under-recognition of this complication in critically ill patients with COVID-19, and proposed endotracheal aspirates be incorporated into the diagnostic algorithm.15 However, as pointed out in response to this, biomarkers, such as galactomannan or Aspergillus PCR, are not validated in tracheal aspirate specimens.16

In view of this, at our institution, we have adopted the following approach for ventilated ICU patients with new or worsening pulmonary infiltrates on chest imaging and laboratory-proven SARS-CoV-2: Screening: Twice-weekly fungal culture and, if feasible, Aspergillus PCR from tracheal aspirate specimens. We do not recommend serum galactomannan as a routine screening test, as the sensitivity was reported to be only 20%.2, 11 Nonetheless, serum galactomannan has a role in predicting prognosis of suspected IPA as emerging data suggest it is indicative of advanced CAPA infection and a positive result is associated with poor outcomes.17 Hence, it can be considered if Aspergillus spp. is cultured or positive by PCR from tracheal aspirates. Antifungal therapy: If patients have persistent fevers for more than 3 days, or recurrence of fevers after at least 48 h of defervescence (despite appropriate antibiotics), with increasing ventilator support, haemoptysis, pleural friction rub or chest pain,11 AND positive Aspergillus spp. culture and/or Aspergillus detected on PCR on two separate occasions from tracheal aspirate specimens, this should then trigger a discussion between the multidisciplinary team of infectious diseases physicians, microbiologists and ICU physicians, to determine if bronchoscopy can be practically and safely performed to confirm further the diagnosis, or if empiric therapy should be commenced should bronchoscopy not be feasible. As per published criteria, a positive Aspergillus spp. culture, Aspergillus PCR or galactomannan ≥1.0 on bronchoalveolar lavage is considered consistent with CAPA in this setting. A serum galactomannan ≥0.5 also supports the diagnosis of CAPA.11

Voriconazole is the recommended first-line treatment for CAPA; however, there are significant drug–drug interactions including interactions with common COVID-19 treatment, such as remdesivir.18 Its narrow therapeutic window and toxicity may worsen the clinical and biochemical status of these ICU patients. Liposomal amphotericin B is the alternative agent, but it is nephrotoxic and is often not suitable as patients already have acute kidney injury. Posaconazole has recently been shown to be non-inferior to voriconazole in the treatment of invasive aspergillosis,19 and it should be considered as the salvage therapy in patients who are not able to tolerate voriconazole or liposomal amphotericin B. Isavuconazole can be considered if prolongation of corrected QT interval (QTc) prolongation is of concern with voriconazole or posaconazole.20

We acknowledged the descriptive nature of our cases and that our case series constitutes a small sample size only. In addition, given that only tracheal aspirates were examined twice weekly, we could only classify these cases as possible CAPA. Nonetheless, we believe that our real-world experience poses several learning points for other institutions across Australia. As demonstrated by Jabeen et al., our experience highlights the challenges and diagnostic dilemma encountered by the clinicians working in environments with varied resources and access to diagnostic tests.

In conclusion, clinicians should be aware of IPA co-infection in ICU patients with COVID-19 infection, particularly those receiving dexamethasone. Given the high mortality, treatment should be considered in deteriorating patients who have new or worsening pulmonary infiltrates and mycological evidence of Aspergillus infection.