ABSTRACT
Background: The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic led to overuse of antimicrobials, which increased concerns regarding antimicrobial resistance. Objective: To measure the impact of a multiplex polymerase chain reaction (PCR) pneumonia panel on empirical antibiotic treatment for patients with critical coronavirus disease 2019 (COVID-19) with suspected bacterial respiratory superinfection. Methods: This descriptive, prospective study was undertaken in a 36-bed intensive care unit from June 2020 to July 2021. Patients with severe COVID-19 who were ventilated and under suspicion of bacterial respiratory superinfection were included in the study. The intervention was a semi-quantitative multiplex PCR alongside concurrent standard cultures. When PCR panel results were expected to be obtained within 3 h of sampling, empirical antibiotic treatment was not administered while awaiting the results. Otherwise, empirical treatment was initiated. Patients classified as 'avoided empirical treatment' avoided 48-72 h of empirical antibiotic therapy. For those patients who received empirical treatment, the PCR panel results were used to decide whether treatment should be escalated, de-escalated, maintained or stopped. Positive and negative predictive values, and 'avoided empirical treatment' were calculated. Medical conduct and panel results were analysed for patients who received empirical treatment. Results: Eighty-two patients (71% male, 29% female) were included in this study. The mean age was 57.5 years, and the mean APACHE II score was 16. Ninety PCR panels were performed, and the negative and positive predictive values were 99.9% and 66.7%, respectively. Empirical treatment was avoided in 61% of episodes. Of those patients who were receiving antibiotics when the PCR panel was performed, treatment was de-escalated in 71%, escalated in 14%, stopped in 9% and maintained in 6%. A diagnosis of bacterial respiratory superinfection was ruled out in 19% of cases. Conclusions: PCR panels prevented the initiation of empirical antibiotic treatment in two-thirds of patients, and led to de-escalation in more than two-thirds of those who had started empirical antibiotic treatment. The high negative predictive value of the PCR panel allowed the diagnosis of bacterial respiratory superinfection to be ruled out. This tool represents a significant contribution to diagnostic stewardship in order to avoid the unnecessary use of antibiotics.
