Mycobiome analyses of critically ill COVID-19 patients

Abstract

Coronavirus disease 2019 (COVID-19)-associated pulmonary aspergillosis (CAPA) is a life-threatening complication in patients with severe COVID-19. Previously, acute respiratory distress syndrome in patients with COVID-19 has been associated with lung fungal dysbiosis, evidenced by reduced microbial diversity and Candida colonization. Increased fungal burden in the lungs of critically ill COVID-19 patients is linked to prolonged mechanical ventilation and increased mortality. However, specific mycobiome signatures associated with severe COVID-19 in the context of survival and antifungal drug prophylaxis have not yet been determined, and such knowledge could have an important impact on treatment. To understand the composition of the respiratory mycobiome in critically ill COVID-19 patients with and without CAPA and the impact of antifungal use in patient outcome, we performed a multinational study of 39 COVID-19 patients in intensive care units (ICUs). Respiratory mycobiome was profiled using internal transcribed spacer 1 sequencing, and Aspergillus fumigatus burden was further validated using quantitative PCR. Fungal communities were investigated using alpha diversity, beta diversity, taxa predominance, and taxa abundances. Respiratory mycobiomes of COVID-19 patients were dominated by Candida and Aspergillus. There was no significant association with corticosteroid use or CAPA diagnosis and respiratory fungal communities. Increased A. fumigatus burden was associated with mortality and, the use of azoles at ICU admission was linked with an absence of A. fumigatus. Our findings suggest that mold-active antifungal treatment at ICU admission may be linked with reduced A. fumigatus-associated mortality in severe COVID-19. However, further studies are warranted on this topic.IMPORTANCEInvasive fungal infections are a serious complication affecting up to a third of patients with severe COVID-19. Nevertheless, our understanding of the fungal communities in the lungs during critically ill COVID-19 remains limited. Evidence suggests a higher fungal burden is associated with prolonged ventilation and higher mortality, although the particular organisms responsible for this link are unclear. Antifungal prophylaxis may be beneficial for reducing the burden of fungal co-infections in COVID-19 intensive care. However, the composition of the fungal microbiome in severe COVID-19 in relation to prophylactic antifungals, as well as how this is associated with survival outcomes, is yet to be studied. Our study provides insights into the lung fungal microbiome in severe COVID-19 and has found antifungal treatment to be associated with lower Aspergillus fumigatus burden and that higher levels of this pathogen are associated with mortality. Therefore, our study suggests mold-active antifungal prophylaxis may be beneficial in severe COVID-19.

Keywords: Aspergillus; CAPA; COVID-19; mycobiome.

Fig 1

Aspergillus and Candida spp. dominate the respiratory mycobiome in critically ill COVID-19 patients. (A) Aspergillus and Candida were the main genera observed in the lungs from COVID-19 patients included in the study (n = 39). Samples are grouped based on CAPA status. (B) A. fumigatus was the predominant species in the mycobiomes of 50% of patients with CAPA, compared to 27% of those without CAPA. (C) Fifty percent of CAPA patients were A. fumigatus positive by species-specific qPCR, compared to 21% of those without CAPA. (D) Alpha diversity measures (observed OTUs and Shannon diversity) trended toward higher diversity in CAPA patients (E) Corticosteroid treatment caused no apparent effect on alpha diversity as measured by observed OTUs or Shannon diversity. (F) A. fumigatus was the predominant species in the mycobiomes of 24% of patients receiving corticosteroids, compared to 22% of those without corticosteroids. Hypothesis testing was applied using Wilcoxon’s rank-sum tests (D,E) or Fisher’s exact tests (B,C,F). Boxplot data represent median and interquartile range. CAPA, COVID-19-associated pulmonary aspergillosis; CS, corticosteroid.

Fig 2

A higher A. fumigatus burden is associated with mortality in critically ill COVID-19 patients. (A) Alpha diversity measures (observed OTUs and Shannon diversity) trended toward higher diversity in critically ill COVID-19 patients who survived. Data represent median and interquartile range. (B) At the genus level, pooled relative abundance mycobiome data from patients who survived (n = 13) indicated a lower proportion of Aspergillus and an apparent increase in the number of observed taxa overall. (C) C. albicans and A. fumigatus were prevalent as the predominant species in a mycobiome in more patients who did not survive (blue, prevalence of 0.44 and 0.32, respectively) than those who did survive (gray, prevalence of 0.32 and 0.08, respectively). Taxa were counted if present at over 50% of the total counts, and only taxa found in at least 10% of the samples of either group are shown. (D) A. fumigatus was the predominant species in 8 of 25 (32%) patients who did not survive, compared to 1 of 13 (8%) of patients who did survive. (E) Analysis using DESeq2 identified A. fumigatus and C. albicans to be at significantly lower abundance in patients who survived. (F) Thirty-two percent of patients who did not survive were A. fumigatus positive by species-specific qPCR, compared to 7% of those which survived. Hypothesis testing was applied using Wilcoxon rank-sum tests (A), Fisher’s exact tests (D and F), or DESeq2 (E). NS, no survival; S, survival.

Fig 3

A. fumigatus is associated with the absence of azole treatment at intensive care unit admission in critically ill COVID-19 patients. Data shown here are for a subset of patients (n = 13) for which azole treatment information was available. (A) Alpha diversity measures (observed OTUs and Shannon diversity) trended higher diversity in COVID-19 patients without azoles at ICU admission. (B) At the genus level, BAL ITS1 mycobiomes from COVID-19 patients who received azoles (n = 5) displayed an absence of Aspergillus, whereas three of eight patients not receiving azoles harbored Aspergillus. For simplicity, genus-level data are shown here. The corresponding species data are shown in Fig. S7A. (C) C. albicans was prevalent as the predominant species of a mycobiome in more patients who received azoles (gray, prevalence of 0.6) than those who did not receive azoles (blue, prevalence of 0.38). A. fumigatus was prevalent as the predominant species in 25% of patients receiving azoles. A. fumigatus was not prevalent in any patients who received azoles. Taxa were counted if present at over 50% of total counts, and only taxa found in at least 10% of samples of either group are shown. (D) A. fumigatus was the predominant species in two of eight (25%) patients who did not survive, compared to none of the patients which did survive. (E) Analysis using DESeq2 identified A. fumigatus to be at significantly lower abundance in patients who received azoles. C. albicans, C. parapsilosis, and C. tropicalis were all at significantly higher abundance in patients receiving azoles. (F) Thirty eight percent (three of eight) of patients who did not receive azoles were A. fumigatus positive by species-specific qPCR, compared to no patients who did receive azoles. (G) All patients receiving azoles were BAL galactomannan negative (ODI of 1 or lower). Two-thirds (four of six) of patients not receiving antifungals were galactomannan positive. Hypothesis testing was applied using Wilcoxon rank-sum tests (A and G), Fisher’s exact tests (D and F), or DESeq2 (E). AZ, azole; ODI, optical density index.

Fig 4

A. fumigatus is associated with mortality in patients with COVID-19 who have not received azole treatment at the intensive care unit admission. Data shown here are for a subset of patients (n = 13) for which azole treatment information was available. (A) When combining the use of azoles at ICU and survival outcomes, Aspergillus was found in the BAL ITS1 mycobiomes from 50% of patients who neither received azoles nor survived (three of six). No Aspergillus was present in samples from the remaining three patient groups. For simplicity, genus-level data are shown here. The corresponding species data are shown in Fig. S7B (B) A. fumigatus levels (measured by qPCR) did not differ significantly between groups (Wilcoxon rank-sum test). However, A. fumigatus burden was observed only in the patient group which neither survived nor received azole treatment at ICU admission. AZ, azole; NS, no survival; S, survival.