Predictors for IAPA and poor outcome (multivariate analysis)
Asthma (OR 12.0 [95% confidence interval (CI) 2.1-67.2]) and days of mechanical ventilation (OR 1.1 [95% CI 1.1–1.2]) were independent predictors for IAPA (figure 2). IAPA (OR 28.8 [95% CI 3.3-253.4]), infection with influenza A (OR 3.3 [95% CI 1.4-7.8]) and illness severity (Simplified Acute Physiology Score, SAPS II) (OR 1.1 [95% CI 1.05-1.10]) were independently associated with poor outcome (figure 2). IAPA was associated with significantly longer median ICU-LOS (29 [95% CI: 11-17] days vs. 5 [95% CI: 4-8] days, p<0.001; figure 3).

Discussion

This study on critically ill influenza patients in seven tertiary care hospitals in Switzerland found an IAPA prevalence of 10.8% over two influenza seasons. IAPA was independently associated with asthma and duration of mechanical ventilation and was an independent risk factor for poor outcome. Other independent predictors of poor outcome were influenza A and higher SAPS II.
IAPA patients needed more and longer organ supportive therapies including mechanical ventilation and vasoactive support and had longer ICU-LOS. Complications and ICU-mortality were more frequent in IAPA.
To our knowledge, asthma was identified as an independent risk factor for IAPA for the first time. This seems biologically plausible. Viral infections like influenza can cause severe exacerbations in patients with asthma and chronic obstructive pulmonary disease (COPD)20,21. Standard treatment of bronchial asthma consists of inhaled and/or systemic corticosteroids and inhaled bronchodilators. Prednisone treatment within 28 days prior to influenza infection has been recognized as a risk factor for developing IAPA1 and corticosteroid treatment is a risk factor for invasive fungal infection in lung disease22. Corticosteroid treatment was also shown to cause a higher fungal burden in the lung23. Therefore, application of inhaled and/or systemic corticosteroids in asthma patients could explain the higher risk for IAPA. Our data also show more frequent corticosteroid treatment in asthma patients during hospitalization. In addition, asthma patients have altered mucociliar clearance of the lung that explains higher rates of fungal growth and colonization in these patients24. High clinical suspicion, early and regular screening for IAPA are therefore warranted in asthma patients. Similarly, this underlines the importance of influenza vaccination for patients with asthma as recommended in Swiss guidelines25.
The colonization with Aspergillus spp. – a prerequisite for the development of IAPA - likely occurs prior to ICU admission as suggested by the POSA-flu trial, where posaconazole prophylaxis started on ICU admission but failed to prevent IAPA or lower mortality in influenza patients15. The authors therefore concluded that development of invasive fungal disease occurs early after influenza infection as 71% of IAPA cases were diagnosed within 24h of ICU admission. This seems plausible since colonization withAspergillus spp. is a known risk factor for developing invasive aspergillosis26. Identification of asthma as a risk factor for IAPA further strengthens this pathophysiologic hypothesis since asthma patients are frequently colonized with Aspergillusspp. 24, 27.
This study identified IAPA, high SAPS II and infection with influenza A as predictors for poor outcome in critically ill patients with influenza. High mortality in patients with IAPA has been reported by us and others1, 14. Interestingly, influenza A was associated with poor outcome that was also shown in a recent meta-analysis of 14 studies of IAPA28. An association of influenza A with an increased risk of bacterial respiratory infections and mortality has been previously suggested29-31, even though it was recently challenged32.
The proportion of IAPA among influenza patients requiring ICU care in Switzerland was similar in the 2017/18 and the 2019/20 seasons14 which were characterized by influenza B with Yamagata (2017/18) predominance vs. similar presence of influenza A(H1N1)pdm09 and B Victoria (2019/20). This is in line with previous reports1, 12 suggesting that IAPA is not restricted to a selected influenza seasons1.
This study is limited by its retrospective design. To optimize pre-analytics and screening of IAPA in influenza patients a screening algorithm was installed at the beginning of the influenza season 2019/20 in all participating ICUs. Despite the recommended screening algorithm for IAPA, sampling of respiratory material and testing for GM was underutilized. This could have resulted in an underestimation of IAPA diagnosis and growth of Aspergillus spp. in respiratory samples in the non-IAPA group. However, growth of Aspergillus spp. was only observed in one patient in the non-IAPA group in which 59% had at least one respiratory sample taken. We therefore assume that most IAPA patients were correctly classified and identified and results can be generalized to critically-ill influenza patients. Because the setting was ICU-specific no conclusions can be made regarding IAPA in patients in an ambulatory setting or hospitalized on the ward. Also, generalizability of results is limited by small numbers of IAPA patients. The observation of a higher risk of IAPA in asthmatic patients does not prove causality and should be confirmed in larger preferably prospective cohort studies.

Interpretation

In conclusion, our data stress the importance of diagnosing IAPA in patients with influenza in the ICU. IAPA was a frequent complication of influenza with high associated mortality, frequent need of organ supportive therapies and longer stay in the ICU. Furthermore, asthma was newly identified as a risk factor for IAPA. We call for increased awareness of IAPA in critically ill asthma patients with influenza, including more intense screening strategies. Prevention efforts through influenza vaccination should be improved in asthma patients as well.