Variability in Durations of Therapy for Gram-Negative Bloodstream Infections and Factors Associated With Prolonged Durations.

Historically, data to inform antibiotic durations for gram-negative bloodstream infections (GN-BSI) were limited. However, 4 randomized controlled trials have recently demonstrated similar outcomes for patients with uncomplicated GN-BSI who received 7 vs 14 days of antibiotics [ 1–5 ]. Adoption of this type of practice-changing evidence can be expedited by antimicrobial stewardship programs [ 6 , 7 ]; yet, an understanding of treatment durations used in clinical practice and factors that influence treatment durations is essential to inform interventions to promote shorter durations. Such information is also critical to determine future areas of research to optimize durations of therapy for GN-BSI [ 8 ]. We aimed to describe treatment durations used for GN-BSI in clinical practice and factors associated with use of prolonged durations.

METHODS

This observational cohort study consisted of consecutive adult patients hospitalized for GN-BSI at 24 US hospitals during 2019 (16 academic, 4 community, 4 Veterans Affairs). The methods have been previously described [ 9 ]. Episodes of GN-BSI due to the following sources were included in this study: urinary, hepatobiliary, gastrointestinal, vascular catheter, respiratory, skin and soft tissue infection, or neutropenic fever without an identified source. Episodes were excluded if death or discharge to hospice occurred within 72 hours of the first positive blood culture collection.

Treatment duration was determined by manual abstraction of the medical record and defined as the total number of calendar days that the patient received inpatient or outpatient antibiotic therapy with in vitro activity against the infecting pathogen. Treatment durations were described for the overall cohort, the subgroup of cases of uncomplicated bacteremia, and the subgroup of cases involving immune compromise. Uncomplicated bacteremia was defined as meeting all the following criteria: the absence of immune compromise; complete source control achieved, if required, within 7 days of the positive blood culture [ 10 ]; and the absence of fever (temperature ≥38 °C) and hypotension (systolic blood pressure <90 mm Hg or receipt of vasopressors) beyond the third calendar day of active antibiotic therapy [ 6 ]. Immune compromise was defined as any of the following: an autoimmune condition with immunosuppressive therapy, active chemotherapy, solid organ or stem cell transplant in the past 12 months, graft-vs-host disease, absolute neutrophil count <500 cells/mL in the 7 days after blood culture collection, or HIV infection with a CD4 cell count <200 cells/mL.

Treatment durations >10 days were categorized as prolonged and ≤10 days as short [ 11–13 ]. Logistic regression was performed to identify variables associated with prolonged treatment durations in the overall cohort and in cases of uncomplicated bacteremia. To avoid confounding from the classification of patients who died early as having received a short duration of therapy, only patients who were alive on day 7 of therapy were included in the models. Variables were selected for inclusion in the models based on prior studies [ 11–16 ] and their clinical relevance as judged by the study team. Final models were developed in compliance with goodness of fit and model complexity according to the Akaike information criterion. The study was approved by the institutional review boards of each participating facility with waivers of informed consent.

RESULTS

A total of 4240 episodes of GN-BSI were included in the analysis. The median age of patients was 67 years (IQR, 55–77); 51% were male ( Table 1 ). Common sources of bacteremia included urinary tract (53%), gastrointestinal (15%), and hepatobiliary (12%). The most frequent pathogens were Escherichia coli (53%), Klebsiella pneumoniae (21%), and Pseudomonas aeruginosa (8%). Half of the episodes were uncomplicated GN-BSI; 30% involved hosts with immune compromise. Overall, 30% required intensive care unit–level care.

The median treatment duration for all GN-BSI episodes was 12 days (IQR, 9–15); 2491 (59%) were treated for >10 days. In the uncomplicated and immune-compromised subgroups, the median treatment duration was 11 days (IQR, 8–14) and 13 days (IQR, 9–16), and 1079 (51%) and 861 (67%) were treated >10 days, respectively. Among all, there was substantial variability in treatment durations at the individual hospital level, ranging from a median of 7 to 14 days ( Figure 1 A ). A similar pattern of variability was observed in the uncomplicated bacteremia and immune compromise subgroups ( Figure 1 B and 1 C ). Of the 24 facilities, 15 (63%) had a median treatment duration >10 days.

In the logistic regression model for the overall cohort ( Supplementary Table 1 ), the following factors were independently associated with a prolonged treatment duration (>10 days): hematologic malignancy, solid organ transplant, isolation of a non-Enterobacterales organism, positive blood culture on day 2 to 5, source control not achieved within 7 days, age ≥65 years, and lack of intravenous-to-oral switch by day 5. In the model including only cases of uncomplicated bacteremia ( Supplementary Table 2 ), isolation of a non-Enterobacterales organism, isolation of a multidrug-resistant organism, and lack of intravenous-to-oral switch by day 5 were independently associated with a prolonged treatment duration.

DISCUSSION

In this large cohort of GN-BSI episodes from 24 US hospitals, 59% were treated with an antibiotic course >10 days, including more than half of those with uncomplicated BSI. There was marked variability in treatment durations across hospitals, overall and for the subgroups with uncomplicated bacteremia and immune compromise.

Randomized trials evaluating different durations of therapy for GN-BSI have focused on uncomplicated bacteremia. Among uncomplicated GN-BSI episodes in this cohort, more than half received an antibiotic course >10 days. These real-world data stand in sharp contrast to a prior survey study in which the majority of physician and pharmacist respondents reported that their preferred duration of therapy was 7 days [ 14 ]. It is worth noting that the first randomized trial demonstrating the effectiveness of 7 days of therapy for uncomplicated GN-BSI was published in late 2018 [ 1 ]; the remaining trials were published after 2019 (the year of cases included in this study) [ 2–4 ]. Thus, it is possible that treatment durations for uncomplicated GN-BSI have since declined; however, there is often a prolonged lag time between the publication and adoption of medical evidence [ 7 ]. Furthermore, reductions in a subset of hospitals in the present study would only increase the variability observed. Overall, our findings suggest that there is likely substantial opportunity to reduce unnecessary variation in treatment durations for uncomplicated GN-BSI by promoting evidence-based 7-day durations of therapy. Given that we found that infections with multidrug-resistant organisms and non-Enterobacterales organisms were independent predictors of prolonged therapy in uncomplicated bacteremia, these may be important areas of focus for antimicrobial stewardship. Since a lack of transition from intravenous to oral therapy was also associated with prolonged durations, encouraging early transition to oral therapy may be a useful intervention to facilitate shorter durations. Individual hospitals have achieved success in shortening treatment durations for such patients [ 17–19 ]; thus, it is essential to disseminate and scale such interventions.

Although reducing treatment durations for uncomplicated GN-BSI represents a clear need, it is important to highlight that half of GN-BSI episodes in this cohort were “complicated,” with nearly a third involving immune compromise. Two-thirds of patients with immune compromise were treated for >10 days. Furthermore, by logistic regression, various forms of immune compromise, including hematologic malignancy and solid organ transplant, were independently associated with prolonged treatment durations. This may in part stem from the fact that patients with severe immune compromise have been excluded from most clinical trials [ 2–4 ], as well as concern for poor outcomes with undertreatment. However, in one randomized trial that included solid organ transplant recipients [ 1 ] and in several observational studies of immune-compromised populations [ 20 , 21 ], the limited available data suggest that outcomes are comparable with shorter and longer durations.

There are several limitations of our study. Despite the broad inclusion of 24 US hospitals, generalizability may be limited. First, all of the hospitals had an active antimicrobial stewardship program, although none had a local guideline for the management of GN-BSI during the study period. One might expect durations of therapy to be even longer in hospitals without a stewardship program. Second, the participating hospitals were disproportionately academic medical centers. Interestingly, we did not observe substantial differences in treatment durations between academic hospitals and community or Veterans Affairs hospitals. Additional limitations include the inability to account for the influence of provider-level characteristics or other facility-level factors that may influence treatment durations. Despite these limitations, this is the largest description of treatment durations for GN-BSI to date and, to our knowledge, the only study evaluating variability in treatment durations and factors associated with prolonged durations.

In summary, treatment durations >10 days for GN-BSI, including uncomplicated bacteremia, are common, while marked variability across hospitals exists. This highlights the need to disseminate and scale successful interventions to reduce inappropriate variability in treatment durations. However, as about half of GN-BSIs are complicated and one-third involve immune compromise, a continued focus of antimicrobial stewardship and clinical trials on uncomplicated GN-BSI alone is insufficient.

Supplementary Material

Contributor Information

Jesse D Sutton, Department of Pharmacy, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, Utah, USA.

Jae Hyoung Lee, Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.

Karen Howard, Department of Pharmacy, University of Utah Health, Salt Lake City, Utah, USA.

Katelyn S Quartuccio, Department of Pharmacy, University of Rochester Medical Center–Highland Hospital, Rochester, New York, USA.

Christopher McCoy, Beth Israel Lahey Health Pharmacy, Boston, Massachusetts, USA.

Alok Gupta, Division of Infectious Diseases, Department of Medicine, University of Rochester Medical Center–Highland Hospital, Rochester, New York, USA.

Judianne C Slish, Department of Pharmacy, University of Rochester Medical Center–Highland Hospital, Rochester, New York, USA.

Drew W Engers, Division of Infectious Diseases, Department of Internal Medicine, Trinity Health Ann Arbor, Ann Arbor, Michigan, USA.

Anurag N Malani, Division of Infectious Diseases, Department of Internal Medicine, Trinity Health Ann Arbor, Ann Arbor, Michigan, USA.

Sara M Karaba, Division of Infectious Diseases, Department of Medicine, School of Medicine, Johns Hopkins University, Baltimore, Maryland, USA.

Kelly E Pillinger, PRIME Education, New York, New York, USA.

Pranita D Tamma, Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA.

Timothy C Jenkins, Division of Infectious Diseases, Department of Medicine, Denver Health, Denver, Colorado, USA; Division of Infectious Diseases, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado, USA.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Note

Potential conflicts of interest . S. M. K. served on an advisory board for Entasis. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

1. Yahav D, Franceschini E, Koppel F, et al. Seven versus 14 days of antibiotic therapy for uncomplicated gram-negative bacteremia: a noninferiority randomized controlled trial. Clin Infect Dis 2018; 69:1091–8.

2. von Dach E, Albrich WC, Brunel A-S, et al. Effect of C-reactive protein–guided antibiotic treatment duration, 7-day treatment, or 14-day treatment on 30-day clinical failure rate in patients with uncomplicated gram-negative bacteremia: a randomized clinical trial. JAMA 2020; 323:2160–9.

3. Molina J, Montero-Mateos E, Praena-Segovia J, et al. Seven-versus 14-day course of antibiotics for the treatment of bloodstream infections by enterobacterales: a randomized, controlled trial. Clin Microbiol Infect 2022; 28:550–7.

4. BALANCE Investigators; Daneman N, Rishu A, Pinto R, et al. Antibiotic treatment for 7 versus 14 days in patients with bloodstream infections. N Engl J Med 2024; 392:1065–78.

5. Turjeman A, von Dach E, Molina J, et al. Duration of antibiotic treatment for gram-negative bacteremia—systematic review and individual participant data (IPD) meta-analysis. EClinicalMedicine 2023; 55:101750.

6. Heil EL, Bork JT, Abbo LM, et al. Optimizing the management of uncomplicated gram-negative bloodstream infections: consensus guidance using a modified Delphi process. Open Forum Infect Dis 2021; 8:ofab434.

7. Morris ZS, Wooding S, Grant J. The answer is 17 years, what is the question: understanding time lags in translational research. J R Soc Med 2011; 104:510–20.

8. Yahav D, Paul M, Van Nieuwkoop C, Huttner A. Is shorter always better? The pros and cons of treating gram-negative bloodstream infections with 7 days of antibiotics. JAC Antimicrob Resist 2022; 4:dlac058.

9. Ince D, Fiawoo S, Choudhury R, et al. Epidemiology of gram-negative bloodstream infections in the United States: results from a cohort of 24 hospitals. Open Forum Infect Dis 2023; 10:ofad265.

10. Engers DW, Tamma PD, Fiawoo S, et al. Transition to oral antibiotic therapy for hospitalized adults with gram-negative bloodstream infections. JAMA Netw Open 2024; 7:e2349864.

11. Isler B, Aslan AT, Benli BS, et al. Duration of antibiotic treatment and timing of oral switching for bloodstream infections: a survey on the practices of infectious diseases and intensive care physicians. Int J Antimicrob Agents 2023; 61:106802.

12. Daneman N, Rishu AH, Xiong W, et al. Duration of antimicrobial treatment for bacteremia in Canadian critically ill patients. Crit Care Med 2016; 44:256–64.

13. Tansarli GS, Andreatos N, Pliakos EE, Mylonakis E. A systematic review and meta-analysis of antibiotic treatment duration for bacteremia due to Enterobacteriaceae. Antimicrob Agents Chemother 2019; 63:e02495-18.

14. Thaden JT, Tamma PD, Pan Q, Doi Y, Daneman N. Survey of infectious diseases providers reveals variability in duration of antibiotic therapy for the treatment of gram-negative bloodstream infections. JAC Antimicrob Resist 2022; 4:dlac005.

15. Daneman N, Shore K, Pinto R, Fowler R. Antibiotic treatment duration for bloodstream infections in critically ill patients: a national survey of Canadian infectious diseases and critical care specialists. Int J Antimicrob Agents 2011; 38:480–5.

16. Corona A, Bertolini G, Ricotta AM, Wilson APR, Singer M. Variability of treatment duration for bacteraemia in the critically ill: a multinational survey. J Antimicrob Chemother 2003; 52:849–52.

17. Erickson RM, Tritle BJ, Spivak ES, Timbrook TT. Impact of an antimicrobial stewardship bundle for uncomplicated gram-negative bacteremia. Open Forum Infect Dis 2019; 6:ofz490.

18. Bavaro DF, Diella L, Belati A, et al. Impact of a multistep bundles intervention in the management and outcome of gram-negative bloodstream infections: a single-center “proof-of-concept” study. Open Forum Infect Dis 2022; 9:ofac488.

19. Deaney MA, Shihadeh KC, Craig A, et al. Standardization in the management of gram-negative bloodstream infections after implementation of a clinical care guideline at a large academic, safety-net institution: a quasi-experimental study. Antimicrob Steward Healthc Epidemiol 2025; 12:e44.

20. Avni-Nachman S, Yahav D, Nesher E, et al. Short versus prolonged antibiotic treatment for complicated urinary tract infection after kidney transplantation. Transpl Int 2021; 34:2686–95.

21. Ranganath N, Yetmar ZA, McCandless AR, et al. Evaluating antimicrobial duration for gram-negative bacteremia in patients with neutropenia due to hematologic malignancy or hematopoietic stem cell transplantation. Transpl Infect Dis 2023; 25:e14085.

Associated Data

Supplementary Materials