You're using an outdated browser. Please upgrade to a modern browser for the best experience.
Please note this is a comparison between Version 3 by Catherine Yang and Version 2 by Catherine Yang.
COVID-19-associated pulmonary aspergillosis (CAPA) is an opportunistic secondary infection, primarily affecting patients in the intensive care unit (ICU) with COVID-19 acute respiratory failure (ARF). Patients affected by COVID-19 ARF and CAPA exhibited increased mortality compared to patients with COVID-19 ARF but without CAPA.
- invasive aspergillosis
- COVID-19
- serum
1. Strengths and Weaknesses of COVID-19 Autopsy Studies for Determining True CAPA Prevalence
1. Strengths and Weaknesses of COVID-19 Autopsy Studies for Determining True CAPA Prevalence
Diagnosis of fungal disease is challenging, especially in the context if ICU. Critically ill patients may present several concomitant infectious and non-infectious processes that may mimic IFI. Accordingly, the only way to achieve a proven aspergillosis classification is to obtain histological diagnosis with evidence of invasive disease [32]. However, lung biopsies are rarely obtained in daily clinical practice and therefore autoptic studies are very important to ascertain the real pathogenetic role of Aspergillus or other fungal disease in this context, as happened in the past. As an example, autoptic studies were very helpful to understand the (poor) pathogenetic role of Candida spp. when isolated in lower respiratory tract of critically ill patients [33].
Post-mortem examination represents the most definitive method to detect invasive mold diseases, including CAPA. Autopsy studies are of particular value for CAPA for mainly two reasons: (1) Invasive pulmonary aspergillosis is generally accompanied by a low ante-mortem detection rate (~40%), and (2) previous autopsy studies suggested ante-mortem overdiagnosis particularly for respiratory tract infections [34,35]. When autopsy is performed with the intention to seek evidence for IFIs, presence of hyphae in histopathological investigation is simple to detect and hence a high likelihood of case identification is warranted. In terms of CAPA, Kula et al. did a systematic review of 50 autopsy studies from 15 different countries, including 677 decedents with severe COVID-19 infection with individual-level data for 443 [36]. Strengths of this review are the high number of investigated cases, performance of standard autopsy (in contrast to minimally invasive techniques) in 82% of decedents, availability (although not utilization) of fungal stain procedures in 91% of decedents, and declaration of important clinical information (e.g., mechanical ventilation status in 81% of decedents). An invasive mold disease was found in only 2% (11/677), with Aspergillus spp. representing the causative pathogen in eight cases, while in two cases mold identification was missing [36]. These findings provide a vigorous counterpart to the reported prevalence in clinical studies, indicating that angioinvasion may only occur in a proportion of cases with airway invasive disease.
Importantly, there are also weaknesses to consider when interpreting evidence from available autopsy studies. First of all, most studies included in the review of Kula et al. were not focused on diagnosis of IFI, because it was largely unknown as a relevant complication of COVID-19 ARF in the early phases of the pandemic when many of those studies were conducted [36]. Consistently, only 38% of autopsy examinations used routine fungal staining of lung tissue. Second, the rate of mechanical ventilated patients was 58%. Therefore, the studies were focused in part in patients with lower risk of CAPA according to recent observations. Third, the number of patients receiving anti-interleukin 6 or corticosteroids was on only 60 (<10%) which is very low in comparison with the current standard of treatment [37]. This may raise the question that some data on the index studies was underreported. Alternatively, this could be secondary to the period of publication of studies included in the review (all in 2020) [36]. In this initial phase of the pandemic, attention on CAPA was poor, and the armamentarium against COVID-19 was weak. It is likely that the prevalence of CAPA may have increased lately, because of the widespread use of immunomodulant therapies. Similarly the study included only few patients with immunocompromising conditions (6%) and this may have reduced the risk for CAPA. We believe that this population is likely to increase in the near future, because of lower response rate of antiSARS-CoV2 vaccine.
Surprisingly, other autopsy studies published later on, and were not included in the aforementioned systematic review, report different findings. In an Italian study reporting results of 45 consecutive autopsies of COVID-19 decedents, proven CAPA was diagnosed in 20% of cases. Of note, fungal staining was routinely performed in that study [38]. In another recent single center autopsy study, six cases of invasive mold diseases (CAPA N = 4, mucormycosis = 2) in eight decedents with severe COVID-19 infection were reported [39].
Diagnosis of fungal disease is challenging, especially in the context if ICU. Critically ill patients may present several concomitant infectious and non-infectious processes that may mimic IFI. Accordingly, the only way to achieve a proven aspergillosis classification is to obtain histological diagnosis with evidence of invasive disease [1]. However, lung biopsies are rarely obtained in daily clinical practice and therefore autoptic studies are very important to ascertain the real pathogenetic role of Aspergillus or other fungal disease in this context, as happened in the past. As an example, autoptic studies were very helpful to understand the (poor) pathogenetic role of Candida spp. when isolated in lower respiratory tract of critically ill patients [2].
2. CAPA as an Important Opportunistic Infection in COVID-19 ICU Patients
Invasive pulmonary aspergillosis has been considered historically a problem in the immunocompromised host. However, in recent years, colonization or infection by Aspergillus spp. in the ICU has been increasingly reported. As an example, in the EPIC II study, an international multicenter study reporting the prevalence of infection in ICUs, reported a prevalence of positive culture of 1.4% in ICU patients with infection [40]. Conversely, a recent study reported a prevalence of 14% of probable pulmonary aspergillosis among ICU patients with ventilator-associated pneumonia [41]. The increased awareness, novel diagnostic methods and change in the population accessing intensive care may explain this wide difference. In COVID-19 ICU patients there are several additional factors that may allow to consider Aspergillus as an important opportunistic pathogen. From a pathophysiological standpoint, viral infections are considered an important trigger for fungal infection by the involvement of a number of immune pathways and mechanisms that may play a crucial role in the co-pathogenesis of viral and fungal lung infections. In fact it has been hypothesized that the damaged epithelium of the airway and suppression of cellular immunity, including defective antigen-specific cytotoxic T lymphocyte responses and impaired phagocyte activities such as phagocytosis, production of cytokines, and reactive oxygen species, formation of neutrophil extracellular traps and killing abilities, are the basis for viral and fungal co-infection. Both damaged epithelium and viral infection is likely to increase IFN production. In addition, IFN-α/β are also produced by alveolar macrophages. This hyperproduction of IFNs may, in turn, suppress monocyte, macrophage and neutrophil recruitment and effector responses that are essential against fungal infections [42].
In several studies assessing risk factors for CAPA among COVID-19 patients, most of attention was reserved for corticosteroids and immunomodulant agents. Corticosteroids are now considered as a backbone treatment for severe or critical COVID-19 infection and the large majority of ICU patients receive or have recently received dexamethasone. Steroids have potent, pleiotropic effects on the immune system that can predispose patients to developing life-threatening invasive aspergillosis. In several studies focused on settings different from COVID-19, corticosteroid use was associated with increased risk of pulmonary aspergillosis [43,44,45]. Studies focused on COVID-19 reported conflicting results. Several papers found that dexamethasone or other corticosteroids were associated with higher risk of CAPA [2,13,46], whereas others failed to find this association [27]. It has to be considered that, as corticosteroids represent a standard treatment for COVID-19, it could be possible that the lack of control without exposure may have biased the analysis in some cases. In a similar fashion, use of interleukin-6 blockers (mainly tocilizumab) was associated with CAPA [2,3].
Post-mortem examination represents the most definitive method to detect invasive mold diseases, including CAPA. Autopsy studies are of particular value for CAPA for mainly two reasons: (1) Invasive pulmonary aspergillosis is generally accompanied by a low ante-mortem detection rate (~40%), and (2) previous autopsy studies suggested ante-mortem overdiagnosis particularly for respiratory tract infections [3][4]. When autopsy is performed with the intention to seek evidence for IFIs, presence of hyphae in histopathological investigation is simple to detect and hence a high likelihood of case identification is warranted. In terms of CAPA, Kula et al. did a systematic review of 50 autopsy studies from 15 different countries, including 677 decedents with severe COVID-19 infection with individual-level data for 443 [5]. Strengths of this review are the high number of investigated cases, performance of standard autopsy (in contrast to minimally invasive techniques) in 82% of decedents, availability (although not utilization) of fungal stain procedures in 91% of decedents, and declaration of important clinical information (e.g., mechanical ventilation status in 81% of decedents). An invasive mold disease was found in only 2% (11/677), with Aspergillus spp. representing the causative pathogen in eight cases, while in two cases mold identification was missing [5]. These findings provide a vigorous counterpart to the reported prevalence in clinical studies, indicating that angioinvasion may only occur in a proportion of cases with airway invasive disease.
Importantly, there are also weaknesses to consider when interpreting evidence from available autopsy studies. First of all, most studies included in the review of Kula et al. were not focused on diagnosis of IFI, because it was largely unknown as a relevant complication of COVID-19 ARF in the early phases of the pandemic when many of those studies were conducted [5]. Consistently, only 38% of autopsy examinations used routine fungal staining of lung tissue. Second, the rate of mechanical ventilated patients was 58%. Therefore, the studies were focused in part in patients with lower risk of CAPA according to recent observations. Third, the number of patients receiving anti-interleukin 6 or corticosteroids was on only 60 (<10%) which is very low in comparison with the current standard of treatment [6]. This may raise the question that some data on the index studies was underreported. Alternatively, this could be secondary to the period of publication of studies included in the review (all in 2020) [5]. In this initial phase of the pandemic, attention on CAPA was poor, and the armamentarium against COVID-19 was weak. It is likely that the prevalence of CAPA may have increased lately, because of the widespread use of immunomodulant therapies. Similarly the study included only few patients with immunocompromising conditions (6%) and this may have reduced the risk for CAPA. This population is likely to increase in the near future, because of lower response rate of antiSARS-CoV2 vaccine.
Surprisingly, other autopsy studies published later on, and were not included in the aforementioned systematic review, report different findings. In an Italian study reporting results of 45 consecutive autopsies of COVID-19 decedents, proven CAPA was diagnosed in 20% of cases. Of note, fungal staining was routinely performed in that study [7]. In another recent single center autopsy study, six cases of invasive mold diseases (CAPA N = 4, mucormycosis = 2) in eight decedents with severe COVID-19 infection were reported [8].
2. CAPA as an Important Opportunistic Infection in COVID-19 ICU Patients
Invasive pulmonary aspergillosis has been considered historically a problem in the immunocompromised host. However, in recent years, colonization or infection by Aspergillus spp. in the ICU has been increasingly reported. As an example, in the EPIC II study, an international multicenter study reporting the prevalence of infection in ICUs, reported a prevalence of positive culture of 1.4% in ICU patients with infection [9]. Conversely, a recent study reported a prevalence of 14% of probable pulmonary aspergillosis among ICU patients with ventilator-associated pneumonia [10]. The increased awareness, novel diagnostic methods and change in the population accessing intensive care may explain this wide difference. In COVID-19 ICU patients there are several additional factors that may allow to consider Aspergillus as an important opportunistic pathogen. From a pathophysiological standpoint, viral infections are considered an important trigger for fungal infection by the involvement of a number of immune pathways and mechanisms that may play a crucial role in the co-pathogenesis of viral and fungal lung infections. In fact it has been hypothesized that the damaged epithelium of the airway and suppression of cellular immunity, including defective antigen-specific cytotoxic T lymphocyte responses and impaired phagocyte activities such as phagocytosis, production of cytokines, and reactive oxygen species, formation of neutrophil extracellular traps and killing abilities, are the basis for viral and fungal co-infection. Both damaged epithelium and viral infection is likely to increase IFN production. In addition, IFN-α/β are also produced by alveolar macrophages. This hyperproduction of IFNs may, in turn, suppress monocyte, macrophage and neutrophil recruitment and effector responses that are essential against fungal infections [11].
In several studies assessing risk factors for CAPA among COVID-19 patients, most of attention was reserved for corticosteroids and immunomodulant agents. Corticosteroids are now considered as a backbone treatment for severe or critical COVID-19 infection and the large majority of ICU patients receive or have recently received dexamethasone. Steroids have potent, pleiotropic effects on the immune system that can predispose patients to developing life-threatening invasive aspergillosis. In several studies focused on settings different from COVID-19, corticosteroid use was associated with increased risk of pulmonary aspergillosis [12][13][14]. Studies focused on COVID-19 reported conflicting results. Several papers found that dexamethasone or other corticosteroids were associated with higher risk of CAPA [15][16][17], whereas others failed to find this association [18]. It has to be considered that, as corticosteroids represent a standard treatment for COVID-19, it could be possible that the lack of control without exposure may have biased the analysis in some cases. In a similar fashion, use of interleukin-6 blockers (mainly tocilizumab) was associated with CAPA [15][19].
3. Diagnosis and Antifungal Treatment of CAPA: How and When?
To date, the largest studies investigating epidemiology of CAPA uniformly report higher mortality rates in patients with CAPA (52%–71%) compared to patients without CAPA (32%–43%) [15][18][19], particularly when both BALF culture and BALF GM are positive [20]. Thus, a high level of alertness and clinical suspicion, optimally resulting in early diagnosis and initiation of antifungal treatment, is vital for sufficient management. Due to the unspecific clinical and radiological findings, mycological evidence represents a main component for establishing a diagnosis. Absence of neutropenia in patients with CAPA results in primarily airway invasive growth (in contrast to primarily angioinvasion in neutropenic patients) and broncho-alveolar lavage (BAL) samples are therefore the preferred sample type for CAPA diagnosis [21]. In contrast, GM from serum is exceedingly lacking sensitivity, indicating that it may only turn positive very late in the disease process once angioinvasion occurs.
According to ECMM/ISHAM consensus criteria, various mycological criteria (i.e., BAL direct examination identifying Aspergillus hyphae, positive Aspergillus spp. culture in BAL, BAL GM of ≥1.0 optical density index (ODI), BAL qPCR of <36 Cq, and serum GM of >0.5) are each sufficient criteria to diagnose CAPA, without one being superior to another [1]. Combination of mycological criteria (e.g., qPCR + GM in BAL) may result in higher specificity [22]. In LMICs where bronchoscopy is scarcely available, validation of upper respiratory tract specimens (e.g., TA) is needed. Recent data suggest high sensitivity of the lateral flow assay (LFA)and GM enzyme linked immunosorbent assay (ELISA) from TA resulting in a high negative predictive value; TA GM testing could represent a potential tool for excluding CAPA, or at least rendering it much less likely. On the flipside, it is important to emphasize that TA GM testing lacks specificity [23][24]. Respiratory specimens, especially BAL for which common diagnostic tests and their respective cut-offs are validated, represent the cornerstone in CAPA diagnosis, in case bronchoscopy is available.
Current guidelines recommend treatment initiation as early as possible in order to hit during the airway invasive phase [1]. This may result in some overtreatment (i.e., treatment in patients that may not develop angioinvasive infection) and adverse drug events, yet there is no viable alternative approach, considering the high mortality rates in patients with CAPA in the absence of antifungal treatment, and those who receive antifungal treatment only once serum GM becomes positive [25]. Voriconazole or isavuconazole represent equal first line drugs, whereas liposomal amphotericin B (L-AMB) is the primary alternative. In suspected azole resistance, voriconazole/isavuconazole can be combined with an echinocandin and L-AMB as an equal alternative in this setting and first choice once resistance is proven [1]. Novel antifungals in the pipeline, namely fosmanogepix and olorofim, may have similar efficacy compared to azoles, yet without the same burden of limiting drug–drug interactions [26]. Hopefully they will overcome the bottleneck of limiting pharmacokinetics and toxicity of current azoles, becoming more favorable treatment options in the near future. While mold active antifungal prophylaxis has been shown to be successful in preventing CAPA in some single center cohort studies [27][28][29], larger level evidence is needed before this approach can be recommended.
3. Diagnosis and Antifungal Treatment of CAPA: How and When?
To date, the largest studies investigating epidemiology of CAPA uniformly report higher mortality rates in patients with CAPA (52%–71%) compared to patients without CAPA (32%–43%) [2,3,27], particularly when both BALF culture and BALF GM are positive [47]. Thus, a high level of alertness and clinical suspicion, optimally resulting in early diagnosis and initiation of antifungal treatment, is vital for sufficient management. Due to the unspecific clinical and radiological findings, mycological evidence represents a main component for establishing a diagnosis. Absence of neutropenia in patients with CAPA results in primarily airway invasive growth (in contrast to primarily angioinvasion in neutropenic patients) and broncho-alveolar lavage (BAL) samples are therefore the preferred sample type for CAPA diagnosis [4]. In contrast, GM from serum is exceedingly lacking sensitivity, indicating that it may only turn positive very late in the disease process once angioinvasion occurs.
According to ECMM/ISHAM consensus criteria, various mycological criteria (i.e., BAL direct examination identifying Aspergillus hyphae, positive Aspergillus spp. culture in BAL, BAL GM of ≥1.0 optical density index (ODI), BAL qPCR of <36 Cq, and serum GM of >0.5) are each sufficient criteria to diagnose CAPA, without one being superior to another [32]. Combination of mycological criteria (e.g., qPCR + GM in BAL) may result in higher specificity [26]. In LMICs where bronchoscopy is scarcely available, validation of upper respiratory tract specimens (e.g., TA) is needed. Recent data suggest high sensitivity of the lateral flow assay (LFA)and GM enzyme linked immunosorbent assay (ELISA) from TA resulting in a high negative predictive value; TA GM testing could represent a potential tool for excluding CAPA, or at least rendering it much less likely. On the flipside, it is important to emphasize that TA GM testing lacks specificity [19,30]. Respiratory specimens, especially BAL for which common diagnostic tests and their respective cut-offs are validated, represent the cornerstone in CAPA diagnosis, in case bronchoscopy is available.
Current guidelines recommend treatment initiation as early as possible in order to hit during the airway invasive phase [32]. This may result in some overtreatment (i.e., treatment in patients that may not develop angioinvasive infection) and adverse drug events, yet there is no viable alternative approach, considering the high mortality rates in patients with CAPA in the absence of antifungal treatment, and those who receive antifungal treatment only once serum GM becomes positive [48]. Voriconazole or isavuconazole represent equal first line drugs, whereas liposomal amphotericin B (L-AMB) is the primary alternative. In suspected azole resistance, voriconazole/isavuconazole can be combined with an echinocandin and L-AMB as an equal alternative in this setting and first choice once resistance is proven [32]. Novel antifungals in the pipeline, namely fosmanogepix and olorofim, may have similar efficacy compared to azoles, yet without the same burden of limiting drug–drug interactions [49]. Hopefully they will overcome the bottleneck of limiting pharmacokinetics and toxicity of current azoles, becoming more favorable treatment options in the near future. While mold active antifungal prophylaxis has been shown to be successful in preventing CAPA in some single center cohort studies [50,51,52], larger level evidence is needed before this approach can be recommended.