COVID-19 associated pulmonary aspergillosis in critically-ill patients: a prospective multicenter study in the era of Delta and Omicron variants

Background During the first COVID-19 pandemic wave, COVID-19-associated pulmonary aspergillosis (CAPA) has been reported in up to 11–28% of critically ill COVID-19 patients and associated with increased mortality. As new SARS-CoV-2 variants emerged, the characteristics of critically ill COVID-19 patients have evolved, particularly in the era of Omicron. The purpose of this study is to investigate the characteristics of CAPA in the era of new variants. Methods This is a prospective multicenter observational cohort study conducted in France in 36 participating intensive care units (ICU), between December 7th, 2021 and April 26th 2023. Diagnosis criteria of CAPA relied on European Confederation of Medical Mycology (ECMM)/International Society for Human & Animal Mycology (ISHAM) consensus criteria. Results 566 patients were included over the study period. The prevalence of CAPA was 5.1% [95% CI 3.4–7.3], and rose to 9.1% among patients who required invasive mechanical ventilation (IMV). Univariable analysis showed that CAPA patients were more frequently immunosuppressed and required more frequently IMV support, vasopressors and renal replacement therapy during ICU stay than non-CAPA patients. SAPS II score at ICU admission, immunosuppression, and a SARS-CoV-2 Delta variant were independently associated with CAPA in multivariable logistic regression analysis. Although CAPA was not significantly associated with day-28 mortality, patients with CAPA experienced a longer duration of mechanical ventilation and ICU stay. Conclusion This study contributes valuable insights into the prevalence, characteristics, and outcomes of CAPA in the era of Delta and Omicron variants. We report a lower prevalence of CAPA (5.1%) among critically-ill COVID-19 patients than previously reported, mainly affecting intubated-patients. Duration of mechanical ventilation and ICU stay were significantly longer in CAPA patients. Supplementary Information The online version contains supplementary material available at 10.1186/s13613-024-01296-0.


Introduction
The societal and individual consequences of pneumonia caused by respiratory viruses, notably due to influenza virus and SARS-CoV-2, are well established.Patients with severe pneumonia may develop acute respiratory failure and require admission to the intensive care unit (ICU).Replication of a respiratory virus in the lower respiratory tract and severe inflammation associated with immune cell infiltration lead to gas exchanges impairment.Viral pneumonia increases patients' susceptibility to bacterial and fungal superinfections, including invasive pulmonary aspergillosis [1,2].Influenza-associated pulmonary aspergillosis (IAPA) has been reported in up to 19-25% of critically ill patients with influenza, associated with poor outcomes [2,3].Coronavirus disease 2019 (COVID-19)-associated pulmonary aspergillosis (CAPA) has similarly emerged as an important coinfection in critically ill patients with COVID-19.The diagnostic criteria for these co-infections combine clinical, radiological, mycological, and histological criteria [4].A recent autopsy study on patients infected with influenza and COVID-19 confirmed the invasive nature of the infection and the similarity of histological lesions observed in both IAPA and CAPA cases [5].A multicenter French study conducted during the first wave revealed that 15% of critically ill patients with COVID-19 requiring invasive mechanical ventilation (IMV) fulfilled the diagnostic criteria for CAPA, which was also associated with poor outcomes [1].In addition to the prognostic impact of CAPA, studies conducted during the first wave identified hostrelated risk factors of CAPA, including age, concomitant treatment with corticosteroids and tocilizumab, and prolonged duration of mechanical ventilation [1,6].
As various SARS-CoV-2 variants have emerged along with the epidemic waves, the characteristics of critically ill COVID-19 patients significantly evolved, especially since the Omicron era: reduced use of IMV, higher rate of immunosuppressed patients, and different treatment approaches as compared to the first COVID-19 wave [7].The purpose of this study is to investigate the characteristics of CAPA with the emergence of Delta variant followed by Omicron and related sublineages, identify potential predictive factors and assess its prognosis impact.

Patients and clinical data
This study is a prospective multicenter observational cohort study.Patients admitted between December 7th, 2021 and April 26th, 2023 in one of the 36 participating ICUs (including 19 from the Greater Paris area) were eligible for inclusion in the SEVARVIR cohort study (NCT05162508) if they presented the following inclusion criteria: age ≥ 18 years, SARS-CoV-2 infection confirmed by a positive reverse transcriptase-polymerase chain reaction (RT-PCR) in nasopharyngeal swab samples, admission in the ICU for acute respiratory failure (i.e., peripheral oxygen saturation (SpO2) ≤ 90% and need for supplemental oxygen or any kind of ventilator support), patient or next of kin informed of study inclusion.Patients with SARS-CoV-2 infection but no acute respiratory failure or with a RT-PCR cycle threshold (Ct) value > 32 in nasopharyngeal swabs were not included.Over the study period, 47% of participating ICUs had rooms with negative pressure, accounting for a total of 37% of the rooms.The study was approved by the Comité de Protection des Personnes Sud-Méditerranée I (N° EudraCT/ID-RCB: 2021-A02914-37).Informed consent was obtained from all patients or their relatives.
Demographics, clinical and laboratory variables were recorded upon ICU admission and during ICU stay.Patients' frailty was assessed using the Clinical Frailty Scale [8].The severity of the disease upon ICU admission was assessed using the World Health Organization (WHO) 10-point ordinal scale [9], the sequential organ failure assessment (SOFA) score [10], and the simplified acute physiology score (SAPS) II score [11].Acute respiratory distress syndrome (ARDS) was defined according to the Berlin definition [12].CAPA diagnosis work-up was at the initiative of the attending clinician in our study (i.e., targeted sampling strategy).The diagnosis and classification (i.e., proven, probable or possible) of CAPA relied on ECMM/ISHAM (European Confederation of Medical Mycology, the International Society for Human and Animal Mycology) international consensus criteria (Additional file 2: Table S1) [4].

SARS-CoV-2 variant determination
Full-length SARS-CoV-2 genomes from all included patients were sequenced by means of next-generation sequencing.Briefly, viral RNA was extracted from nasopharyngeal swabs in viral transport medium using NucliSENS ® easyMAG kit on EMAG device (bioMérieux, Marcy-l'Étoile, France).Sequencing was performed with the Illumina ® COVIDSeq Test (Illumina, San Diego, California), which uses 98-target multiplex amplifications along the full SARS-CoV-2 genome.The libraries were sequenced with NextSeq 500/550 High Output Kit v2.5 (75 Cycles) on a NextSeq 500 device (Illumina).The sequences were demultiplexed and assembled as full-length genomes by means of the DRAGEN COVID-Seq Test Pipeline on a local DRAGEN server (Illumina).Lineages and clades were interpreted using Pangolin and NextClade, before being submitted to the GISAID international database (https:// www.gisaid.org).

Statistical analyses
Descriptive results are presented as means (± standard deviation [SD]) or medians (1st-3rd quartiles) for continuous variables, and as numbers with percentages for categorical variables.Two-tailed p values < 0.05 were considered statistically significant.Unadjusted comparisons according to CAPA status (CAPA patients vs. non CAPA patients) were performed using Chi-square or Fisher's exact tests for categorical variables, and t-tests or Mann-Whitney tests for continuous variables, as appropriate.
Multivariable logistic regression models were performed to identify the parameters most associated with CAPA, entering variables associated with a p-value < 0.20 in univariable analysis and those previously shown to be potential confounding factors, including age and gender, then applying a stepwise backward approach by retaining only variables statistically significant at a relaxed p < 0.10 level.Adjusted odds ratios (aORs) along with their 95% confidence intervals (CI) were computed.
To assess the potential effect of CAPA occurrence on subsequent prognosis, 90-day overall survival was estimated using the Simon-Makuch method [13] and was compared using the Mantel-Byar test between those patients having developed a CAPA and those who had not, considering CAPA occurrence as a time-dependent variable.Cox proportional hazards regression modelling was used to compute Hazard ratios (HR) and their corresponding CIs, with CAPA as a time-dependent covariate and further adjusting for important prognostic factors (i.e., age, gender, baseline SOFA score and immunosuppressive status).Landmark analyses of 90-day overall survival by CAPA status were also conducted as sensitivity analyses.To do so, participants who died or were censored before a 5-days landmark point (as the median time of CAPA occurrence) were excluded from the landmark analyses at day 5 of, allowing a better control of the so called 'immortal-time bias' (i.e., patients dying early in the study have a limited time to develop CAPA thus guaranteeing poorer outcomes in the patients unexposed to CAPA) and yielding potentially more accurate results by increasing the number of CAPA patients at risk when starting at a 5-day time point compared to earlier time points when the number of CAPA patients at risk are usually smaller.
An exploratory unsupervised clustering analysis was achieved allowing to explore the heterogeneity of the population using the Kohonen's self-organized map (SOM) methodology [14], allowing us to build 2-dimensional maps from multidimensional datasets.In a nutshell, each map is divided into districts in which patients are located by the SOM algorithm on the basis of their characteristics: patients with similar features are closely located on the maps, while patients with distinct profiles are farther from each other, hence allowing to identify key differences or similarities among them by drawing visual comparisons of unique or overlapping patient characteristics and disease subtypes.Clinical or biological variables considered as relevant were included in this analysis.The SOMs were obtained with the Numero package framework for the R statistical platform [15] after principal component analysis adapted for mixtures of qualitative and quantitative variables was applied (PCAMix) [16,17].
All measurements were taken from distinct samples.Variables with missing information used for the evaluation of the risk factors of CAPA by logistic regression, for the evaluation of the effect of CAPA occurrence on overall survival by Cox proportional hazards regression modelling and for the exploratory clustering analysis by self-organizing maps, were imputed using the k-nearest neighbors (k-NN) approach.Analyses were performed using Stata V16.1 statistical software (StataCorp, College Station, TX, USA), and R 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria).

Population
Over the study period, a total of 566 patients were admitted in one of the 36 participating ICUs and included in the study, including 242 patients requiring IMV during ICU stay.Twenty-nine patients (5.1% [95% CI 3.4-7.3])fulfilled the diagnosis criteria for CAPA (Fig. 1).The prevalence of CAPA was higher in patients who required IMV (N = 22/242, 9.1%) than in those who did not (N = 7/324, 2.2%).

Baseline characteristics
Compared to non-CAPA patients, those who fulfilled CAPA diagnosis criteria, did not show statistically significant differences regarding age, gender and comorbidities, except for more frequent immunosuppression (N = 15/29, 54% vs. N = 174/537, 34%, p = 0.03) (Table 1).Although the proportion of vaccinated patients did not significantly differ according to CAPA status, CAPA patients had more frequently negative anti-S SARS-CoV-2 IgG antibodies than non-CAPA patients, possibly linked to the higher proportion of immunosuppressed patients in the former than in the latter group.The median delay between the first symptoms of disease and ICU admission was significantly longer in CAPA than in non-CAPA patients (8 [6-11] vs. 7 [3][4][5][6][7][8][9][10] days; p = 0.03).The viral load of SARS-CoV-2 in the upper respiratory tract (estimated with the cycle threshold of RT-PCR) did not significantly differ between groups (Table 1).There was no significant difference according to CAPA status regarding the severity of the disease at ICU admission, as reflected by the SOFA and SAPS II scores and the WHO 10-point ordinal scale (Table 1).
The Simon-Makuch estimates of overall survival from ICU admission to day-90 is depicted in Fig.

Characteristics of CAPA patients
According to CMM/ISHAM CAPA definitions, 24 patients (83%) fulfilled the criteria of proven/probable CAPA and 5 patients (17%) were classified as possible CAPA (Table 3).The detailed diagnostic criteria are shown in Additional file 1: Figure S1.The diagnosis of CAPA was made a median of 5 [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16] days after ICU admission and 6 [2][3][4][5][6][7][8][9][10][11][12][13] days after tracheal intubation (among the CAPA patients requiring IMV).Half of the patients were immunosuppressed: 10 (66%) had an oncohematological malignancy, and 4 (27%) had received an organ transplant.Among the 29 CAPA patients, 28 (97%) received an antifungal treatment during ICU stay.Voriconazole and isavuconazole were the two most frequently administered antifungal drugs (Table 3).To better characterize the phenotypic differences between CAPA and non-CAPA patients, an exploratory analysis using the SOM method was performed to plot 2-D maps of patients grouped according to their characteristics (Fig. 3).SOM analysis depicted the observed differences according to CAPA status.As shown in the figure, CAPA patients tended to cluster in the upper left area of the map, where the highest frequencies of immunosuppression, the highest values of the SOFA and SAPS II scores also clustered.Patients with the highest rates of day-28 mortality and use of IMV during ICU stay clustered in the same area of the map than CAPA patients.

Factors associated with CAPA
In multivariable analysis after missing data imputation, four factors associated with higher risk of CAPA were retained after stepwise analysis: increased SAPS II score (aOR 1.03 [95% CI 1.003-1.05],p = 0.028), immunosuppression (aOR 2.65 [1.13-6.20],p = 0.025), a SARS-CoV-2 Delta variant (aOR 2.72 [1.12-6.58],p = 0.027), and to a lower extent an increased delay between the first symptoms and ICU admission (aOR 1.03 [0.997-1.06],p = 0.077) (Table 4).Age and IMV support during ICU P = 0.943 P = 0.926 Fig. 2 a Simon-Makuch estimates of overall survival status from ICU admission to day-90 according to CAPA status; b Simon-Makuch estimates of overall survival status from ICU admission to day-90 according to CAPA status using landmark times at 5 days Liposomal amphotericin B 2 (7) stay were not significantly associated with CAPA at the p < 0.10 level.Results of the multivariable analysis on raw data are presented in Additional file 3: Table S2.

Discussion
To the best of our knowledge, we herein report the largest cohort study investigating the prevalence and the characteristics of CAPA among critically-ill COVID-19 patients in the era of Delta and Omicron SARS-CoV-2 variants.The main results of our study are the following: (i) the prevalence of CAPA in the whole cohort was 5.1%, and 9.1% among patients requiring IMV; (ii) CAPA patients were more frequently immunosuppressed, required more frequently IMV, vasopressors and renal replacement therapy during ICU stay; (iii) CAPA was not associated with day-28 mortality, but with a longer duration of mechanical ventilation support and ICU stay; and (iv) SAPS II score at ICU admission and the delay between the first symptoms and ICU admission were independently associated with CAPA.
The prevalence of CAPA herein reported (5.1%) is lower than previously reported in studies conducted during the first wave in mechanically ventilated ICU patients: 11-28% [1,6,18,19] or in more recent studies: 16-33% [20,21].Several factors may explain these results.First, these studies were carried out only in mechanically ventilated patients, while more than half of the patients in our cohort did not require IMV.Indeed, in the subgroup of patients requiring IMV, the reported prevalence was higher (9.1%).Second, these studies routinely screened for diagnosis criteria for CAPA once or twice a week, whereas it was at the initiative of the attending clinician in our study (i.e., targeted sampling).Whether bronchoalveolar lavage sampling should be routinely performed to increase the sensitivity of the diagnosis criteria has been suggested [22].We acknowledge that our study may be associated with a lower prevalence of CAPA than that reported in other studies performing routine screening [20], however our targeted sampling strategy reflects real-life practice in a nation-wide study.The most appropriate diagnostic strategy for CAPA remains to be defined.Indeed, as most of the CAPA diagnostic criteria are nonspecific (i.e., biomarkers, non-specific clinical and/ or radiological signs), a systematic diagnostic approach (as opposed to targeted sampling) might be associated with a lower pre-test probability and thus a lower positive predictive value of CAPA diagnosis.Third and importantly, the clinical phenotype of critically-ill COVID-19 patients has evolved in line with the natural course of the disease, with older and frailer patients, more frequently immunosuppressed, and presenting with higher severity scores at ICU admission [7].Yet, in the current series, the prevalence of CAPA among patients infected with Omicron (4.8%) was lower than that observed in patients infected with Delta (9.2%), and lower than that reported during the first pandemic wave [1].These findings might point to a variantrelated effect.CAPA has been associated with impaired antifungal immunity (i.e., altered integrity of the epithelial barrier, and decreased capacity to phagocytise and kill Aspergillus spores and to destroy Aspergillus hyphae) [23].Recent findings suggested a reduced evasion of variant Omicron from innate immunity [24,25], as compared to pre-existing variants, which might be associated with more effective antifungal immunity and hence a lower prevalence of CAPA.Another potential factor might be the inherent disorganization during the For instance, immunosuppressed patients were more frequently located in the upper districts and also had higher serum urea levels, less frequent Delta variant infection, higher SAPS II and SOFA scores and day-28 mortality rates.WHO World Health Organization, SOFA Sequential Organ Failure Assessment, SAPS II Simplified Acute Physiology Score II, MV mechanical ventilation first epidemic wave, resulting in less frequent use of rooms with negative pressure in comparison with the period of this study (i.e., Delta and Omicron era).We also describe the existence of CAPA in nonintubated patients, although occurring in only 2.2% of patients, which had rarely been reported to date.Indeed, previous studies investigating CAPA included almost exclusively patients requiring IMV [1,18,20,21,26].However, two multicenter studies reported that a minority of CAPA occurred in non-intubated patients [6,27], 6% and 12% of CAPA patients, respectively.Unfortunately, their design did not allow for a reliable assessment of the prevalence of CAPA in this setting.In our study, however, most of non-intubated CAPA patients had a specific underlying condition (five patients were immunocompromised (four haematological malignancies and one solid organ transplantation) and one patient had a chronic cavitary pulmonary aspergillosis, rather than a "classical" CAPA.No classical risk factor was identified for the last patient.Therefore, it appears appropriate to screen non-intubated COVID-19 patients for CAPA in case of underlying immunosuppressive status or other risk factors, rather than routinely.
Previous studies identified various factors associated with CAPA: age, long-term corticosteroids use, chronic obstructive pulmonary disease, haematological malignancy, IMV and its prolonged duration, tocilizumab treatment, especially in association with corticosteroids [1,6,18,19,[28][29][30][31].In our study, we could identify three factors independently associated with CAPA in multivariable analysis, including SAPS II score at admission, immunosuppression and a SARS-CoV-2 Delta variant.The first two factors are thus in line with previous findings; whereas the Delta variant effect has not yet been described [20].It is important to note most of the previous studies included patients before the Omicron era, and our study is the largest one investigating CAPA in the era of Omicron.
We did not observe an association between day-28 mortality and CAPA status in our study, in contrast with previous studies that reported a high mortality associated with CAPA [1,6,18,19,26,32].Nevertheless, duration of mechanical ventilation and ICU stay were significantly longer in CAPA patients than non-CAPA patients.Moreover, vital status was only reported at day-28, and it is likely that CAPA status was associated with longer-term prognosis.In line with previous studies, the diagnosis of CAPA was mostly reliant on serum and/or respiratory fungal markers.Thus, the prognostic relevance of CAPA per se could be questioned: it might be considered a simple severity indicator rather than a truly invasive super-infection.A recent pathology-based studies has highlighted the invasive nature of CAPA [5], thereby supporting systematic CAPA screening of intubated patients and specific antifungal treatment.
Our study has some limitations.We were unable to compare these results with those from the first waves, as this prospective cohort began during Delta variant era.The relatively small number of CAPA patients included may have limited our statistical ability to show between-group differences.Non-systematic screening for CAPA may have underestimated the prevalence of CAPA.However, our study also has strengths, including the constitution of a unique prospective multicenter cohort reporting the prevalence and characteristics of CAPA in the era of Delta and Omicron SARS-CoV-2 variants.

Conclusion
To conclude, we report a lower prevalence of CAPA (5.1%) among critically-ill COVID-19 patients in the era of Delta and Omicron variants than previously reported, and mainly affecting intubated-patients.SAPS II score at ICU admission, immunosuppression status and a SARS-CoV-2 Delta variant were independently associated with CAPA.Even though CAPA was not associated with day-28 mortality, duration of mechanical ventilation and ICU stay were significantly longer in CAPA patients.

Fig. 3
Fig. 3 Unsupervised analysis of the clinical and biological characteristics of the by self-organized maps (SOMs).Unsupervised analysis by SOM automatically located patients with similar clinical and paraclinical parameters within 1 of 40 small groupings ("districts") throughout the map.The more similar the patients, the closer on the map.Each individual map shows the mean values or proportions per district for each characteristic: blue indicates the lowest average values, red the highest, with numbers shown for a selection of representative districts in each SOM.For instance, immunosuppressed patients were more frequently located in the upper districts and also had higher serum urea levels, less frequent Delta variant infection, higher SAPS II and SOFA scores and day-28 mortality rates.WHO World Health Organization, SOFA Sequential Organ Failure Assessment, SAPS II Simplified Acute Physiology Score II, MV mechanical ventilation

Table 1 (
continued) CAPA COVID-19-associated pulmonary aspergillosis, ICU intensive care unit, Ct cycle threshold, WHO World Health Organization, SOFA Sequential Organ Failure Assessment, SAPS II Simplified Acute Physiology Score II, NIV non-invasive ventilation, C-PAP continuous-positive airway pressure, MV mechanical ventilation, ECMO -values come from unadjusted comparisons using Chi-square or Fisher's exact tests for categorical variables, and t-tests or Mann-Whitney tests for continuous variables, as appropriate.No adjustment for multiple comparisons was performed.Bolded p-values are significant at the p < 0.05 level

Table 2
Management and outcomes of critically ill COVID-19 patients (n = 566) during their intensive care unit stay according to Coronavirus disease (COVID-19)-associated pulmonary aspergillosis (CAPA) status

Table 4
Predictors of CAPA occurrence by univariable and multivariable logistic regression models in critically ill patients with COVID-19: results after missing data imputation (n = 566) aOR (CI 95%): adjusted Odds Ratio (95% confidence interval) CAPA COVID-19-associated pulmonary aspergillosis, ICU intensive care unit, SAPS simplified acute physiology score; SOFA Sequential Organ Failure Assessment p-values come from multivariable logistic regression models Bolded p-values are significant at the p < 0.05 level