For a study, researchers sought to investigate patterns of co-infection and intensive care unit-acquired infections (ICU-AI) and the link to corticosteroid administration. This multinational, multicenter observational study comprised adult patients with PCR-confirmed COVID-19 diagnoses who were admitted to ICUs during the peak of COVID-19 wave 1 (February 15th to May 15th, 2020). Co-infections at ICU admission, infections contracted there, infections with multidrug resistant organisms (MDRO), and antibiotic use were all covered by the data. Comparing continuous variables and frequencies was done using the Mann-Whitney U test and Pearson’s Chi-squared, respectively. The “complete” matching method of the R library MatchIT was used to perform propensity score matching for variables related to ICU-acquired infections. Data on 4,994 patients were available. About 716 people (14%), 85% of whom were previously taking antibiotics, were admitted with co-infections caused by bacteria. ICU-AI was created in 2,715. Bacterial pneumonia was the most frequent ICU-AI (44% of infections), followed by fungal pneumonia (9% of patients), and MDRO (25%) infections. Antimicrobial exposure was higher in ICU patients with infections compared to those without. Pre-pandemic surveillance reports were significantly outnumbered by ICU-AI occurrences per 1,000 ICU days. Different ICUs used corticosteroids in different ways. A univariate analysis revealed that ICU-AI occurred in 43% of patients who weren’t receiving corticosteroids and in 58% of patients who were. In the propensity-matched group, ICU-AI was present in 71% of patients getting corticosteroids and 52% of patients not receiving corticosteroids after correcting for potential confounders. Both an MDRO infection and ICU-AI were linked to the length of corticosteroid therapy. Co-infections during ICU admission were exceedingly uncommon in patients with severe COVID-19 in the first wave, but antibiotic use was significantly higher than recommended. Corticosteroid use was frequently associated with ICU-AI.

Source: ccforum.biomedcentral.com/articles/10.1186/s13054-022-04108-8