key: cord-0826139-eqxujy5j
authors: Borman, Andrew M.; Palmer, Michael D.; Fraser, Mark; Patterson, Zoe; Mann, Ciara; Oliver, Debra; Linton, Christopher J.; Gough, Martin; Brown, Phillipa; Dzietczyk, Agnieszka; Hedley, Michelle; McLachlan, Sue; King, Julie; Johnson, Elizabeth M.
title: COVID-19-Associated Invasive Aspergillosis: Data from the UK National Mycology Reference Laboratory
date: 2020-12-17
journal: J Clin Microbiol
DOI: 10.1128/jcm.02136-20
sha: 43e7fb2ad33ba00c1b4bed16cd9c5f49aedc8288
doc_id: 826139
cord_uid: eqxujy5j

COVID-19-associated pulmonary aspergillosis (CAPA) was recently reported as a potential infective complication affecting critically ill patients with acute respiratory distress syndrome following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, with incidence rates varying from 8 to 33% depending on the study. However, definitive diagnosis of CAPA is challenging. Standardized diagnostic algorithms and definitions are lacking, clinicians are reticent to perform aerosol-generating bronchoalveolar lavages for galactomannan testing and microscopic and cultural examination, and questions surround the diagnostic sensitivity of different serum biomarkers. Between 11 March and 14 July 2020, the UK National Mycology Reference Laboratory received 1,267 serum and respiratory samples from 719 critically ill UK patients with COVID-19 and suspected pulmonary aspergillosis. The laboratory also received 46 isolates of Aspergillus fumigatus from COVID-19 patients (including three that exhibited environmental triazole resistance). Diagnostic tests performed included 1,000 (1-3)-β-d-glucan and 516 galactomannan tests on serum samples. The results of this extensive testing are presented here. For a subset of 61 patients, respiratory specimens (bronchoalveolar lavage specimens, tracheal aspirates, and sputum samples) in addition to serum samples were submitted and subjected to galactomannan testing, Aspergillus-specific PCR, and microscopy and culture. The incidence of probable/proven and possible CAPA in this subset of patients was approximately 5% and 15%, respectively. Overall, our results highlight the challenges in biomarker-driven diagnosis of CAPA, especially when only limited clinical samples are available for testing, and the importance of a multimodal diagnostic approach involving regular and repeat testing of both serum and respiratory samples.

I nvasive aspergillosis (IA) complicating severe influenza pneumonia is well described from anecdotal case reports, and more recently from cross-center and large cohort studies (1) (2) (3) (4) (5) . Of particular note, mortality was significantly higher in patients with influenza/aspergillosis than in those with influenza alone (2, 4) . There exists significant variation in the reported incidence of IA in critically ill influenza patients, with incidence rates between 7.2 and 28.1% reported from different studies (1) (2) (3) (4) (5) . These variations in calculated incidence are likely to reflect a combination of real regional differences in Aspergillus predominance/exposure, variable access to different diagnostic testing regimens, different clinical approaches to managing influenza patients in intensive care units (ICUs) and diagnosing IA, and variations in awareness of influenza-associated pulmonary aspergillosis (IAPA) (5) . In addition, direct comparisons of the incidence have been hindered by the employment of nonstandardized reference definitions and diagnostic algorithms for IAPA, since the European Organization for Research and Treatment of Cancer Mycoses (EORTC/MSG) criteria for the diagnosis of IA in immunocompromised patients are not directly applicable to nonneutropenic patients in ICUs (5) (6) (7) . In an attempt to circumvent these limitations in existing reference definitions, modified diagnostic algorithms have been proposed for IA in nonimmunocompromised ICU patients (2, 5, 6, 8, 9) , including the AspICU definition (6) , which showed promising sensitivity and specificity for critically ill ICU patients in whom Aspergillus spp. had been recovered from respiratory specimens (7) . Based on current evidence, galactomannan (GM) testing on serum samples lacks sensitivity compared to testing with fluid recovered from bronchoalveolar lavage (BAL), serum (1-3)-␤-D-glucan (BDG) testing lacks specificity, and there is currently insufficient evidence to support the role of Aspergillus lateral flow devices (LFD) and PCR on serum samples in the diagnosis of IAPA (6, 7) . Thus, IAPA classification frequently relies at least in part upon a positive BAL GM test (7) .

Since late 2019, a novel Betacoronavirus (severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2]) has rapidly spread around the world, causing a global pandemic of coronavirus-19 disease (COVID-19) with over 13 million cases and in excess of 500,000 deaths to date (14 July 2020). The disease spectrum in COVID-19 ranges from asymptomatic infection to severe illness with acute respiratory distress syndrome (ARDS) requiring critical care, with 20 to 40% of hospitalized symptomatic patients developing ARDS (10) (11) (12) . An increasing number of anecdotal case reports and case series have identified COVID-19-associated pulmonary aspergillosis (CAPA) as an additional infective complication of COVID-19 (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) . Depending on the study and the region of origin, reported incidence rates for CAPA ranged from approximately 5% (14) to 20 to 35% (17) (18) (19) (20) (21) . Similar to IAPA, many questions remain concerning the diagnostic criteria used to define CAPA (12) , and diagnosis remains challenging. Radiological findings on COVID-19 and IA in nonneutropenic patients are nonspecific and show considerable overlap. Moreover, GM testing of BAL fluids is often unavailable due to the restricted role recommended for bronchoscopy in COVID-19 and local reluctance to perform bronchoalveolar lavages due to the aerosol-generating nature of this intervention (12, 23, 24) . Currently, the diagnostic value of serum GM and BDG testing or Aspergillus LFDs remains largely unknown in CAPA, although initial studies suggest that GM/BDG might lack sensitivity in this context (12, 15, 19, 23) . While upper respiratory tract specimens (tracheal aspirates [TA], nonbronchoscopic lavage [NBL] specimens) are more readily available than BAL fluids from intubated COVID-19 patients, GM testing is not validated on such specimens, and isolation of an Aspergillus sp. in the absence of other positive biomarkers should be interpreted with a degree of caution (23) . Finally, due to fears surrounding potential aerosol generation during autopsy, postmortem proof of fungal infection to substantiate putative cases of CAPA is absent in all but very rare reports (22) . Due to the above limitations, the diagnosis of a substantial proportion of the putative cases of CAPA reported to date (10/35) has been based solely upon recovery of Aspergillus spp. from respiratory secretions, with no substantiating evidence from additional positive biomarkers (reviewed in reference 12). Similar to IAPA, variations in diagnostic algorithms employed in different centers are likely to contribute to the widely variable incidence rates reported for CAPA to date.

As described abroad (18) , due to ease of access to samples, many centers in the United Kingdom have adopted a weekly or twice weekly screening approach using serum BDG and GM testing, either alone or in conjunction with culture of upper respiratory tract specimens and Aspergillus-specific PCR, to diagnose CAPA in intubated/ventilated COVID-19 ICU patients. Between 11 March and 14 July 2020, during the first "peak" of SARS-CoV-2 infections in the United Kingdom, the Public Health England UK National Mycology Reference Laboratory (MRL) received a substantial number of specimens from critically ill COVID-19 patients across the entire United Kingdom, including over 1,000 serum samples and about 90 BAL, NBL, and TA samples from over possible (putative) and proven CAPA if, in addition, the patients had one or more than one, respectively, of the following: positive Aspergillus PCR test on respiratory secretions or blood, recovery of Aspergillus spp. after culture of respiratory samples, positive GM test on BAL fluid or serum, positive BDG test on serum, or positive LFD test on serum or BAL fluid. However, since multiple and repeated tests were performed on many of the patients described in the current study, a degree of common sense was also applied to rule out likely erroneous results; for example, a single positive BDG test that was flanked by multiple negative tests in the same patient in the absence of any other mycological or biomarker evidence of infection despite multiple tests would not be considered sufficient to change the case definition (see, for example, patients 8, 49, and 52 [ Table 1 ]). Since all current definitions of proven IAPA/CAPA require histopathological evidence of infection, which is difficult to obtain in critically ill COVID patients, none of the cases described in this study could be categorized as proven CAPA.

Between 11 March and 14 July 2020, the MRL received 1,267 samples from 719 critically ill UK patients with COVID-19, including 1,178 serum samples, and 89 respiratory samples. From a total of 1,000 BDG antigen tests performed, 818 (81.8%) were negative (BDG concentration, Ͻ80 pg/ml), with 655 (65.5%) BDG concentrations below the lower limit of detection of the assay (Ͻ30 pg/ml). BDG concentrations in positive serum samples were 80 to 249 pg/ml in 130 samples (13.0%), 250 to 500 pg/ml in 30 samples (3.0%), and Ͼ500 pg/ml in 22 serum samples (2.2%). Similarly, the vast majority (508/516; 98.4%) of serum samples that were subjected to GM antigen testing were negative (index value, Ͻ0.5). Multiple samples and specimen types were available for a subset of the 719 patients for whom samples were submitted, allowing a rudimentary analysis of which biomarkers might have better power in diagnosing CAPA. GM) had no evidence of invasive fungal infection (no positive biomarkers). In patients with a single fungal positive biomarker, that biomarker was most likely to be BDG antigen (83/97; 85.6%), with 30% (3/10) of patients having a positive BAL GM antigen test as the sole positive biomarker and no patients having a positive serum GM antigen test as the single positive biomarker. Unsurprisingly, the relative proportions of serum BDG and serum and BAL GM tests that were positive were substantially increased in patients with multiple positive fungal biomarkers ( Table 2) . Table 3 summarizes the results of biomarker tests and mycological examinations for 15 selected patients with possible (putative), probable, or proven CAPA, stratified according to recent guidelines for IAPA and aspergillosis in nonimmunocompromised ICU patients (6, 7) . Patients were selected for inclusion on the basis of multiple positive biomarker(s) in serum and/or respiratory specimens. Serum BDG tests were positive at least once in 12/15 patients with suspected CAPA, whereas serum GM testing appeared less sensitive (positivity in 5/15 patients). BAL fluids were submitted to the MRL for 5/15 patients with suspected CAPA listed in Table 3 . Filamentous fungal elements were seen in direct microscopy and A. fumigatus recovered in culture in two successive BAL specimens from a single patient (patient 1), who was positive by GM, Aspergillus-specific PCR, and Aspergillus LFD with BAL fluid, but serum BDG and serum GM negative. BAL GM tests were positive in 4/5 patients for whom BAL fluid was available for testing. Finally, Aspergillus-specific PCR was positive on serum or BAL fluid in 3/15 patients, and the Aspergillus-specific LFD test was strongly positive or positive in 3/15 patient samples and weakly positive in a further 6 (total positivity, 9/15 cases). Overall, the results of fungal biomarker testing and mycological examination of samples submitted from these 15 patients with likely CAPA underscore the importance of employing a multifaceted approach, with no single biomarker or even a 2-biomarker combination being capable of detecting all potential cases. Since histopathological proof of infection was not available, no cases fulfilled current AspICU guidelines for proven CAPA, despite some patients having multiple positive biomarkers plus recovery of Aspergillus spp. from respiratory secretions.

The MRL also received 46 isolates of A. fumigatus, 3 of which were resistant to the This patient is case 1 in Table 3 , who was also AspLFD strongly positive.

Journal of Clinical Microbiology triazole antifungals and had mutations consistent with resistance development due to environmental exposure (i.e., the presence of the typical tandem repeat and single nucleotide polymorphisms in the cyp51A gene were detected using the AsperGenius PCR assay), 2 isolates of Aspergillus niger, and 1 isolate of Rhizopus arrhizus, all of which had been cultured from respiratory secretions from COVID-19 patients across the United Kingdom (Table 4) . According to the clinical information submitted with isolates in conjunction with additional biomarker testing that was requested for some of the patients from which organisms were isolated, at least 6/46 isolates of A. fumigatus had been recovered from patients who fulfilled the diagnostic criteria for probable/proven CAPA (Table 4 ). Insufficient clinical information was available for all other isolates to be able to make a judgment on clinical significance. While the data presented in the above sections provide a preliminary insight into the diagnostic utility of different fungal biomarkers in patients suspected of having CAPA, they do not contribute to our understanding of the potential prevalence of CAPA, since the full panel of samples with diagnostic utility (repeat serum samples plus BAL fluid/other respiratory secretions) was not submitted to the MRL for the vast majority of patients. In an attempt to begin to address this issue, we examined a subset of 61 patients who at least partly fulfilled the criteria of serum plus respiratory samples. Table 1 summarizes the results of mycological examination and biomarker testing for these 61 patients. Aspergillus fumigatus was recovered in culture from only 2 BAL specimens out of the 83 respiratory samples analyzed (case 4 in Table 1 ϭ patient 1 in Table 3 ). Of the 61 patients, 13 (21.3%) had at least 1 positive fungal biomarker (possible CAPA), with only 2/61 patients (3.3%) having 2 or more independent positive biomarkers (probable) and 1/61 patients (1.6%) with multiple positive biomarkers and recovery of an Aspergillus sp. from BAL fluid. Taken together, these data suggest that approximately 5% of patients had probable/proven CAPA, with a further ϳ15% having possible CAPA.

Here, we have presented the results of fungal biomarker testing on serum and respiratory samples and mycological examination of respiratory secretions performed on samples from over 700 critically ill COVID-19 patients across the United Kingdom, including a subset of 61 patients for which multiple sample types were available. In the majority of patients, there was no mycological evidence of CAPA (all fungal biomarkers were negative and respiratory samples failed to yield growth of Aspergillus spp.). The data obtained with samples from patients with possible/probable and proven CAPA (Tables 1 and 3 ) allow a preliminary and somewhat crude analysis of the diagnostic value of individual biomarkers and testing combinations. Measurement of serum BDG antigen levels appears potentially promising in detecting CAPA in that this biomarker was positive in 13/15 patients with possible/probable/proven CAPA (Table 3 ) and appears to be more sensitive than serum GM testing. In addition to CAPA, candidemia is another potential complication of severe COVID, and the MRL received 25 isolates of yeast from confirmed cases of candidemia (recovery of the organism from blood cultures) in ICU patients diagnosed with COVID-19. Serum BDG was elevated in all four of the patients from whom we received serum samples (data not shown). However, none of the patients listed in Tables 1 and 3 had any evidence of invasive or disseminated Candida infections, ruling out the possibility that elevated BDG was due to candidemia in those patients. However, in general, BDG positivity is not specific for CAPA in critically ill patients with COVID. Similar to BDG antigen testing on serum, GM antigen tests on BAL fluids (where samples were available) were positive in 4/5 of the patients with probable/proven CAPA who had multiple positive biomarkers (Table 3) but were positive in only 5 of 12 patients with possible/probable/proven CAPA when the 61 patients with both serum and respiratory samples available were analyzed (Table 1) .

It is possible that discrepant serum biomarkers (negative serum BDG/GM) in the Table 2 in addition to the following: NBL, nonbronchoscopic lavage; azole R , pan-triazole resistant isolates; vasc. line tip, vascular line tip.

single patient who was BAL GM positive and microscopy and culture positive is an indicator of the general lack of sensitivity of serum biomarkers. However, we cannot rule out that these results are an indication of heavy colonization of damaged pulmonary epithelium with Aspergillus sp. hyphae prior to the development of active invasive fungal disease. Indeed, distinguishing between Aspergillus colonization as opposed to invasive infection is likely to be difficult in this patient population and is likely to be further confounded in those COVID-19 patients with preexisting chronic pulmonary disease (32) . Moreover, as has been suggested previously (12, 23, 32) , recovery of Aspergillus spp. in culture from respiratory secretions in COVID-19 patients should be interpreted with caution given the ubiquitous nature of A. fumigatus spores. The MRL always performs fluorescence microscopy for the mycological examination of all respiratory samples to aid in the interpretation of the results of conventional culture. Here, we have described a safe method for the microscopic examination of respiratory secretions from COVID-19 positive patients to aid regional laboratories in diagnostic decision making. Of concern is that 3 of the 46 isolates (6.5%) of A. fumigatus from respiratory secretions of COVID-19 patients were resistant to azole antifungal agents. The presence of the mutations associated with emergent resistance due to environmental exposure to agricultural azoles was confirmed by the AsperGenius assay. This is a concerning finding, as it rules out the use of voriconazole as first-line therapy in these patients and highlights the necessity of susceptibility testing of all isolates from suspected infections.

The results from the current study suggest that the incidence of probable/proven and possible CAPA is approximately 5% and 15%, respectively. However, there are a number of limitations to this study. It is not a case series that describes successive patients admitted to a single institution, which would be better suited to addressing incidence rates for CAPA. In addition, since samples were submitted to the MRL for analysis, accurately establishing a denominator to aid incidence calculations is impossible due to two confounding scenarios. On the one hand, it is likely that samples were preferentially submitted on certain patients that local clinicians strongly suspected of having CAPA. Conversely, for a large number of patients (especially those where only weekly BDG testing was requested), it is likely that sampling formed part of a routine surveillance program in long-stay ICU patients. Further limitations include (i) the possibility of missing or incomplete data since the MRL relied upon information that accompanied the samples that was supplied by the referring laboratory, (ii) the fact that MRL had no control over sample types/frequency of sampling in hospitals and microbiology laboratories nationwide, with the result that there were insufficient or inappropriate samples for comprehensive diagnostic testing for many patients, and (iii) the fact that with very few exceptions, information on antifungal therapy and clinical outcome was missing. However, despite these limitations, the current study highlights the risk of invasive pulmonary aspergillosis in critically ill COVID-19 patients and adds to the growing literature concerning CAPA.

Finally, a recent study failed to find histological evidence of CAPA in postmortem biopsy samples from six patients diagnosed with probable CAPA on the basis of positive GM tests on BAL fluids with or without recovery of Aspergillus spp. in culture (33) . All 6 patients had received combination antifungal therapy with voriconazole plus anidulafungin following the diagnosis of probable CAPA. However, given the short duration of their disease and treatment, it would seem unlikely that negative histological results reflected clearance of genuine fungal infections. Rather, such reports lend support to suggestions that a more stringent classification may be required for CAPA cases, compared to existing ones which were developed for IAPA (32, 34) . At the very least, the present study underscores the challenges surrounding the diagnosis of CAPA and the importance of the serial evaluation of multiple fungal biomarkers in both serum and respiratory samples in addition to conventional mycological examination of respiratory samples as part of a comprehensive multimodal approach (26, (32) (33) (34) . The precise determination of the best testing algorithm to accurately diagnose CAPA will require further studies. However, from the results presented here, we would agree with the recent suggestion (35) that in addition to radiological imaging, serial screening for CAPA in ICU patients with deteriorating respiratory function should include (i) regular (at least weekly) Aspergillus antigen testing of serum samples, (ii) regular (at least weekly) BDG testing of serum samples, (iii) Aspergillus antigen testing of BAL fluids (where available) or nondirected lavages/tracheal aspirates, and (iv) Aspergillus PCR in conjunction with conventional mycological examination (microscopy and culture) of respiratory secretions if available.

High rates of influenza-associated invasive pulmonary aspergillosis may not be universal: a retrospective cohort study from Alberta

Invasive aspergillosis in patients admitted to the intensive care unit with severe influenza: a retrospective cohort study

Invasive pulmonary aspergillosis is a frequent complication of critically ill H1N1 patients: a retrospective study

Higher mortality of severe influenza patients with probable aspergillosis than those with and without other coinfections

Review of influenza-associated pulmonary aspergillosis in ICU patients and proposal for a case definition: an expert opinion

AspICU Study Investigators. 2012. A clinical algorithm to diagnose invasive pulmonary aspergillosis in critically ill patients

Performance of existing definitions and tests for the diagnosis of invasive aspergillosis in critically ill, adult patients: a systematic review with qualitative evidence synthesis

Invasive pulmonary aspergillosis in patients with chronic obstructive pulmonary disease

Clinical relevance of Aspergillus isolation from respiratory tract samples in critically ill patients

Full spectrum of COVID-19 severity still being depicted

Risk factors associated with acute respiratory distress syndrome and death in patients with coronavirus disease

COVID-19 associated pulmonary aspergillosis (CAPA): from immunology to treatment

Fatal invasive aspergillosis and coronavirus disease in an immunocompetent patient

Clinical characteristics of invasive pulmonary aspergillosis in patients with COVID-19 in Zhejiang, China: a retrospective case series

Invasive pulmonary aspergillosis complicating SARS-CoV-2 pneumonia: a diagnostic challenge

Invasive pulmonary aspergillosis complicating COVID-19 in the ICU: a case report

Prevalence of putative invasive pulmonary aspergillosis in critically ill patients with COVID-19

COVID-19-associated invasive pulmonary aspergillosis

Shimabukuro-Vornhagen A. 2020. COVID-19 associated pulmonary aspergillosis

Incidence of invasive pulmonary aspergillosis among critically ill COVID-19 patients

Confirmed invasive pulmonary aspergillosis and COVID-19: the value of postmortem findings to support antemortem management

Diagnosing COVID-19-associated pulmonary aspergillosis

American Association for Bronchology and Interventional Pulmonology (AABIP) statement on the use of bronchoscopy and respiratory specimen collection in patients with suspected or confirmed COVID-19 infection

A prospective study of fungal biomarkers to improve management of invasive fungal diseases in a mixed specialty critical care unit

Galactomannan testing and Aspergillus PCR in same-day bronchoalveolar lavage and blood samples for diagnosis of invasive aspergillosis

The evolution and evaluation of a whole blood polymerase chain reaction assay for the detection of invasive aspergillosis in hematology patients in a routine clinical setting

European Aspergillus PCR Initiative (EAPCRI). 2011. Evaluation of Aspergillus PCR protocols for testing serum specimens

Lack of standardization in the procedures for mycological examination of sputum samples from CF patients: a possible cause for variations in the prevalence of filamentous fungi

Is the COVID-19 pandemic a good time to include Aspergillus molecular detection to categorize aspergillosis in ICU patients? A monocentric experience

Detection of invasive pulmonary aspergillosis in critically ill patients by combined use of conventional culture, galactomannan, 1-3-beta-D-glucan and Aspergillus specific nested polymerase chain reaction in a prospective pilot study

Late histopathologic characteristics of critically ill COVID-19 patients: different phenotypes without evidence of invasive aspergillosis, a case series

Fungal infection during COVID-19: does Aspergillus mean secondary invasive aspergillosis?

Confronting and mitigating the risk of COVID-19

We are grateful to the clinicians and microbiologists from across the United Kingdom for submitting patient samples to the MRL for analysis.