Diagnosis of Invasive Fungal Disease in Coronavirus Disease 2019

Approaches and Pitfalls

P. Lewis White


Curr Opin Infect Dis. 2021;34(6):573-580. 

In This Article

Diagnosis of Coronavirus Disease 2019-associated Pulmonary Aspergillosis

A range of testing options are available for the diagnosis of invasive aspergillosis but determining accurate test performance for the diagnosis of CAPA is difficult when results are used to classify the entity. When comparing test positivity for a range of assays across both blood and respiratory samples, it is clear that no single test detects all cases. Collating the data from 68 cases of CAPA described in six studies published early in to the course of the pandemic but with CAPA redefined using a single classification is shown in Table 2.[10,20–24] It confirms no single test generates sensitivity close to 100%, highlighting the potential need for combined testing. Positivity rates are greater when testing respiratory samples, with galactomannan enzyme immunoassay (GM-EIA) and Aspergillus PCR providing the greatest sensitivity. Interestingly, rates of Aspergillus cultured from the respiratory tract were also moderate but could reflect the recovery of Aspergillus from the upper respiratory tract, which while confirming the presence of Aspergillus within the patient is not necessarily specific to disease and should be supported with mycological positivity in samples from deeper within the respiratory tract or in blood. Positivity in blood samples is generally lower, reflecting limited invasion by Aspergillus in COVID-19 patients, who commonly lack the host factors considered to impart risk for invasive aspergillosis. However, serum (1–3)-β-D-Glucan (BDG) positivity appears to provide greater sensitivity over other blood biomarkers. Given the broad fungal detection range of BDG but lack of capacity to differentiate different IFD and numerous sources of BDG false positivity, combining BDG testing with other mycological tests is paramount.[10,25,26] A recent prospective, multicentre evaluation by the European Confederation of Medical Mycology (ECMM) confirmed the findings in Table 1 when testing up to 109 patients with CAPA defined using recent international consensus definitions.[27,28] Seventy-seven percent of CAPA cases were positive (index ≥1.0) by GM-EIA on broncholalveolar lavage (BAL) fluid, compared with 73% by Aspergillus PCR, with Aspergillus being cultured from the respiratory tract in up to 62% of case-based samples. The culture of Aspergillus is also pivotal to performing azole susceptibility testing, with cases of azole-resistant CAPA documented.[29] Serum galactomannan was only positive (index ≥0.5) in 19% of cases, but specificity of GM-EIA was excellent in both serum (99.5%) and BAL fluid (97.6%).[27] The use of Aspergillus lateral flow assays may enhance access to antigen testing of both serum and BAL fluid, whereas performance outside the COVID-19 cohort is similar to GM-EIA testing validation for CAPA is currently limited.[28]

The median time to CAPA presentation is 10 days (range 0–51 days) post-ICU admission, highlighting the need for prolonged and frequent mycological testing to ensure an earlier diagnosis.[9] Bartoletti et al.[24] demonstrated that although 47% of CAPA patients had GM-EIA positivity in BAL fluid within the first 2 days of admission, the majority demonstrated positivity over a longer period (>5 days). The subsequent prolonged testing period questions the suitability of BAL sampling that is invasive to the patient and raises infection control concerns in the COVID-19 patient. Testing nondirected bronchial lavage (NBL) fluid is a possible alternative respiratory sample to BAL fluid, requiring less invasive sampling using a closed suction catheter that minimizes infection control risks. The sensitivity/specificity of GM-EIA testing of NBL was 86 and 95%, respectively, with higher index greater than 4.5 values increasing specificity further (99%), performance, which is comparable with GM-EIA testing of BAL fluid in the non-COVID-19 critical-care patient.[10,30,31] Recommended thresholds for determining GM-EIA positivity in NBL are currently higher than those for BAL fluid, highlighting the uncertainty in specificity associated with NBL testing.[28]

The incidence of CAPA is obviously dependent on the diagnostic strategy applied, and significant variation in incidence has been reported.[8] Classification based on single positive mycology results will likely generate higher incidences but the confidence in classification will vary considerably dependent on the source of positivity and its subsequent signal strength. Recovery of Aspergillus spp. from the upper respiratory tract may indicate airway contamination/colonization but should be used as a trigger for a diagnostic work-up.[32] Although obtaining consecutive positive upper respiratory tract cultures does increase confidence in a diagnosis of CAPA, it is far from conclusive.[32] Positivity in lower respiratory samples increases the likelihood of CAPA but false-positive GM-EIA results can occur in BAL fluid. Ideally GM-EIA BAL fluid positivity should be supported with additional mycological evidence, although GM-EIA specificity is proportional to galactomannan index value.[31,33] High index values on initial GM-EIA testing have also been associated with a poor patient prognosis.[24] Although GM-EIA and Aspergillus PCR positivity in blood of the COVID-19 patient is generally limited, it is likely more specific for CAPA.[32] Serum BDG positivity requires aetiological specific support to overcome the issues discussed above.

In an attempt to standardize the classification of CAPA, various diagnostic strategies have been proposed.[7,10,26,28,32,34] In a recent evaluation of CAPA evidence, a diagnostic work-up, based primarily on bronchoscopy and BAL fluid testing is recommended for all mechanically ventilated COVID-19 patients with unexplained respiratory deterioration or a positive Aspergillus culture from the respiratory tract.[32] Performing bronchoscopy also permits visualization of plaques/eschars associated with Aspergillus tracheobronchitis that may occur in the COVID-19 patient.[28,32] Undoubtedly, BAL fluid is the primary sample for the diagnosis of CAPA and the authors provide a solid basis for performing bronchoscopy, but pragmatically obtaining these samples during the peaks of the COVID-19 pandemic will be difficult, particularly in resource-limited settings. False-positive BAL fluid results will also occur and performing multiple tests (Microscopy/Culture/GM-EIA/Aspergillus PCR) is recommended, with multiple positive tests enhancing confidence in the CAPA diagnosis, something supported by an earlier expert opinion paper.[34]s Although screening of serum with GM-EIA and BDG is not recommended because of the potential for low sensitivity, it is very difficult to facilitate screening over the required, prolonged period on the basis of BAL fluid testing. Evidence above and derived/amended from the studies included in the taskforce report demonstrates that serum-BDG sensitivity at 47% is similar to respiratory culture at 45% (currently proposed as trigger point for work-up) and may warrant BDG inclusion as a trigger alongside the testing of more easily obtainable respiratory samples.[32] Persistent serum BDG positivity and/or mycological evidence in the non-BAL respiratory samples would trigger a diagnostic work-up, including bronchoscopy and further blood biomarkers to confirm a diagnosis.

The ECMM/ISHAM consensus CAPA definitions confirm the preference towards testing BAL fluid but incorporate the testing of more easily obtainable respiratory samples (e.g. NBL) and adjust classification accordingly.[28] As with all classifications proven disease is based on positive histology/microscopy/culture from a tissue biopsy, rarely obtained ante-mortem. Autopsy evidence of CAPA has provided low rates of confirmation, with a recent review confirming IFD in only 2% of deceased COVID-19 patients.[35] However, this could be indicative of limited tissue and angio-invasion in the CAPA patient, although a recent autopsy study did provide high rates (20%) of proven CAPA.[36] It is also important to remember that histological evidence is highly specific for confirming disease but sensitivity is insufficient to exclude it. In two studies, radiology typical of invasive aspergillosis was visualized in approximately 50% of CAPA patients and evidence of cavitation or well defined nodular lesions on CT should heighten the suspicion of CAPA, leading to diagnostic work-up but CT alone is not sufficient to confirm or refute CAPA.[10,26,28,32] Whenever present, radiology typical of invasive aspergillosis may provide clinical evidence sufficient to weight classifications so that lesser mycological evidence is required to define CAPA, compared with patients with nonspecific chest radiology.[10] It is important to remember that the ECMM/ISHAM CAPA definitions have been developed in response to urgent clinical need and through international consensus on the current information available for the diagnosis of IA, much of which has been gained outside the COVID-19 patient.[28] Both the incorporation of NBL testing and the current exclusion of upper respiratory culture positivity have been questioned by different groups.[37,38] The opposing views expressed, not only highlights the diagnostic dilemma encountered by centres working under different clinical pressures, including limited resources but also indicate that the current ECMM/ISHAM CAPA definitions provide a well balanced and solid platform to base studies, while awaiting further evidence required to redefine the definitions.[28,39]