Baseline predictors of antibiotics de-escalation from empirical therapies in an intensive care unit: a five-year retrospective study

Background: The alarming rate of bacterial resistance has urged for vigorous antibiotic de-escalation cultures worldwide. However, the art of de-escalation remains challenging as most clinicians have the fear or stigma that antibiotic de-escalation is strongly associated with higher mortality. This study aimed to determine the factors strongly correlated with higher mortality rates after antibiotic de-escalation and possibly serve as a benchmark study for developing a validated antibiotic de-escalation tool in the near future. 

Method: This retrospective study was conducted on patients hospitalized in a medical intensive care unit of a referral tertiary care center and subjected to antibiotics de-escalation over five years. Independent factors associated with mortality post antibiotics de-escalation were assessed. Various data were collected, including patient demographics, admission and discharge dates, a primary source of infection and free of infection status, choices of antibiotic used and de-escalation history, vital signs, data on X-ray changes, relevant laboratory investigations, microbiological culture history, mortality status, history of COVID-19 infection, presence of central line, number of vasopressors used, ventilator settings and respective SOFA scores. 

Results: Prevalence of mortality among patients continued on broad-spectrum antibiotics empirically is remarkable. Through this study, it was found that the significant factors associated with mortality post antibiotic de-escalation were hospital-acquired infection (HAI) (Adjusted OR: 12.56; 95% CI: 2.88,54.98; p < 0.001), systolic blood pressure in mmHg (Adjusted OR: 0.08; 95% CI: 0.02,0.33; p < 0.001), heart rate in bpm (Adjusted OR: 0.12; 95% CI: 0.03,0.45; p:0.002), SOFA scores (Adjusted OR: 21.44; 95% CI: 3.55,129.52; p:0.001) and number of vasopressors (Adjusted OR: 38.46; 95% CI: 1.54,959.30; p:0.026). 

Conclusion: De-escalation of antibiotics is considerable as the identified significant factors serve as baseline predictors which would certainly be helpful especially during the decision-making process of antibiotic de-escalation among the clinicians. Therefore, antibiotics de-escalation is highly encouraged to reduce the mortality risks.