Diagnostic dilemma in COVID-19-associated pulmonary aspergillosis.

We thank the community for rapid recognition and characterisation of COVID-19-associated pulmonary aspergillosis (CAPA), increasingly observed in people with severe SARS-CoV-2 infection. Consistent with previous efforts to standardise definitions for invasive fungal infections,1, 2, 3 Philipp Koehler and colleagues propose new definitions for CAPA and provide management recommendations. We have questions and concerns for feedback.

Although we agree that standardised definitions are necessary to facilitate enrolment into clinical trials, consensus definitions for invasive fungal infections were historically supported by observational studies and not intended to guide clinical care.1, 2 As a newly recognised syndrome, we worry that the proposed definitions for CAPA are not adequately supported by evidence; premature confidence in definitions risks biasing outcomes of future research and directing inappropriate management, without first establishing a requisite level of evidence.

The proposed CAPA definitions are highly reliant on bronchoscopy, which is variably used, especially in surge conditions with strained infection control. A bronchoscopy-driven approach will inevitably underestimate the burden of CAPA and potentially skew trial enrolment towards people with more invasive disease.

Also, we have concerns with the biomarker cutoffs proposed. Whereas investigators have done detailed studies to determine cutoffs for galactomannan enzyme immunoassays using appropriate measures (ie, receiver operating characteristic curves) in other populations,3, 5 we are not aware of similar data to support recommendations for positivity at the multiple levels proposed by Koehler and colleagues, combined with requirements for repeated testing. This expert proposal is particularly problematic when cutoffs are not aligned with local regulatory recommendations. Should clinicians and investigators ignore their regulatory-cleared biomarker cutoffs in the absence of supportive evidence? Moreover, Koehler and colleagues proposed various cutoffs for non-bronchoalveolar lavage biomarkers, which have not been cleared or validated. Although this proposal is reasonable to generate data in early research settings, one can hardly say that we have enough evidence to derive consensus.

Multiple issues arise with using definitions in different clinical and research contexts. For instance, conservative definitions are not actionable for clinical care or prevention studies—settings where the earliest therapy is essential to improve clinical outcomes. We thank our colleagues for early efforts to understand and define this new entity, but fear that more caution is needed to acknowledge critical gaps in data. We believe that establishing consensus definitions for CAPA requires more efforts, especially those directed towards deriving biomarker performance characteristics.