Intensive care unit-acquired infections more common in patients with COVID-19 than with influenza – Nature.com

In this Swedish retrospective cohort study, mechanically ventilated patients with COVID-19 experienced a higher incidence of ICU-acquired infections compared to those with influenza. Staphylococcus aureus was identified as the most common pathogen causing VA-LRTI among patients with influenza and COVID-19, while gram-negative bacteria as a group caused the majority of VA-LRTI in patients with COVID-19. We found an association between ICU-AI and increased risk of mortality in patients treated with corticosteroids. Our data further suggest that corticosteroid treatment in COVID-19 is a risk factor for acquiring secondary bacterial infections in the ICU.

The differing risk of ICU-AI in patients with COVID-19 as opposed to influenza accords with other studies2,11,21,22,23,24. It may be explained by factors such as increased demand on the healthcare system during the COVID-19 pandemic11,25, alterations of immune responses caused by SARS-CoV-221, a high proportion of ARDS in COVID-19, more frequent prone positioning23, and prolonged IMV and ICU stays11,26. Although we noted no difference in ICU LoS between the COVID-19 and influenza cohorts, there was a small difference in time on IMV. Consistent with findings from other studies23,26, more males were observed in critical COVID-19 cases than in influenza cases. This may account for the different incidence rates, as this and other studies suggest that male gender is a risk factor for ICU-AI15,27.

There was no significant difference in the percentage of patients with corticosteroid treatment between the two cohorts. However, the indication for corticosteroid treatment to patients with influenza was airway obstruction and/or sepsis with lower doses and shorter duration than recommended in severe COVID-19. Furthermore, antibiotic treatment on admission has been shown to be a risk factor for ICU-AI2,28,29, and early initiation of antibiotics was high throughout the pandemic, despite the low frequency of co-infections on admission in patients with COVID-19. On the other hand, it is possible that the lower incidence of ICU-AI in the influenza cohort is partly explained by earlier diagnosis and targeted treatment of co-infection, while some co-infections in the COVID-19 cohort might been missed initially and later misinterpreted as ICU-AI.

As the pandemic developed, incidence rates of ICU-AI in patients with COVID-19 increased. A similar pattern, but with slightly lower incidence rates, was seen in a recent Swedish study on VA-LRTI29. The differing incidence rates of ICU-AI during the pandemic can be partly explained by a shift in corticosteroid treatment, for as our study and several others have suggested, corticosteroid treatment is a risk factor for ICU-AI2,15,22,29,30. Moreover, later in the pandemic patients were more critically ill and had more co-infections on admission, possibly affecting the risk of ICU-AI. Nor can we rule out other variables, such as changes in management or staffing at the ICU31, different SARS-CoV-2 strains, or vaccinations32, any of which may have affected the risk of ICU-AI throughout the pandemic.

Other studies have demonstrated the same association between ICU LoS and IMV duration, while reports on mortality are conflicting15,24,29,30,33. Our findings demonstrate an increased risk of mortality with ICU-AI in patients with corticosteroid treatment as compared to patients who have not received corticosteroids. This may in part reflect the higher mortality that occurred in later waves in contrast to the first. Although glucocorticoids have been shown to reduce mortality12,34, later studies have indicated that not all patients with severe COVID-19 may benefit from corticosteroid treatment15,35,36. We did not find any interaction between age and corticosteroid treatment on the risk of ICU-AI, but it cannot be ruled out that certain patient categories might be affected differently by corticosteroid treatment. Further riskbenefit studies of the association between corticosteroid treatment, ICU-AI, and outcome in hospitalized patients are needed.

The microbial pattern we observed in VA-LRTI is consistent with that seen elsewhere11,14,22,29,30. Although we found a larger discrepancy between the two cohorts than other studies observed11,23,24,37, this may have been due to the small number of patients with influenza and ICU-AI. A shift in the microbial pattern was observed between early and late VA-LRTI, with an increase in more difficult-to-treat microbes in later stages, consistent with findings reported in other studies11,29,30. Possible explanations for this are alterations in lung microbiota38, increase of biofilm-active bacteria39, as well as an overuse of antibiotics2. We noted a change throughout the pandemic towards more broad-spectrum antibiotic treatment on admission in patients with COVID-19. Broad-spectrum antibiotics are a risk factor for ICU-AI28 and may possibly facilitate the development of more complicated infections. Although the rate of MDRO was comparatively low23,40, there is a risk of decreasing antibiotic susceptibility with the overuse of antibiotics41,42.

The major strengths of our study are the large sample size of patients on IMV due to COVID-19 and our detailed examination of the medical charts for each case. There are however some important limitations to consider: First, the retrospective nature of the study. Second, the small comparison group, due to the relatively few patients on IMV as a result of influenza, especially during the COVID-19 pandemic. The inclusion period for the two cohorts also differed somewhat, possibly affecting the prevalence of MDRO. Third, most patients receiving corticosteroid treatment were hospitalized after the first wave, so it is possible that there were coinciding changes in management that further affected the risk of ICU-AI. Fourth, most samples from the lower respiratory tract were not taken with protected brush. This may have resulted in some colonization cultures and contaminations being included for analysis.

Continue reading here:

Intensive care unit-acquired infections more common in patients with COVID-19 than with influenza - Nature.com