Abstract

Introduction Bloodstream infections occur commonly in the United States and are associated with significant morbidity and mortality. Bacteremia caused by gram-positive organisms has established guidelines for the use of follow-up blood cultures (FUBCs), whereas the use of FUBCs in bacteremia due to gram-negative bacilli (GNBs) does not. One study showed that positive FUBCs were more common in gram-positive coccus bacteremia (21%) than in GNB bacteremia (6%), calling into question the overall clinical utility of using FUBCs in GNB bacteremia. We investigated the factors associated with positive FUBCs in patients with GNB bacteremia in our institution to better guide clinical practice. Methods We used a case-control design for this study. We included patients 18 years or older admitted to our hospitals between January 1, 2014, and December 31, 2018, for a duration of ≥24 hours and who had a positive blood culture (BC) for GNB. Cases were defined as any patient with a positive FUBC drawn ≥24 hours after the initial positive BC. Control subjects were defined as patients with a negative FUBC after an initial positive BC. Characteristics of cases and control subjects were compared using descriptive statistics. A multivariable regression model using stepwise variable selection was used to identify factors significantly associated with having a positive FUBC. Results We identified 70 cases meeting the inclusion criteria. We compared these 70 cases with 528 control subjects. The cases and control subjects did not differ significantly in age, sex, race, or other predefined comorbid conditions. Significant predictors for having a positive FUBC included (1) organism resistance to initial antibiotic therapy (odds ratio [OR], 3.26; P = 0.035), (2) glomerular filtration rate less than 30 mL/min (OR, 2.35; P = 0.010), (3) initial admission to critical care unit (OR, 1.72; P = 0.049), (4) central line–associated infection (OR, 3.99; P = 0.004), (5) Klebsiella group infection (OR, 2.18; P = 0.015), and (6) Serratia marcescens infection (OR, 3.53; P = 0.037). Initial discordant BC was a negative predictive factor (OR, 0.44; P = 0.018). Follow-up blood culture drawn more than 48 hours after the initial positive BC had a lower likelihood of being positive than those drawn 24 to 48 hours after the initial BC (OR, 0.59; P = 0.037). Discussion We identified 6 independent risk factors for positive FUBCs as delineated previously. These factors provide guidance for further research and for future guideline development. The timing of the FUBCs had a significant impact on positivity. Follow-up blood cultures in GNB bacteremia should be obtained more than 48 hours after the initial positive BC.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call