Optimal Duration of Antibiotic Treatment for Gram-Negative Bacteremia

In patients with uncomplicated gram-negative bacteremia, are 7 days of antibiotic therapy as effective as 14 days?

Clinicians should treat patients with uncomplicated gram-negative bacteremia with antibiotic therapy for 7 days. (Strength of Recommendation: A, systematic reviews and meta-analyses of randomized controlled trials [RCTs].) In comparing 7 vs 14 days of therapy, there are no differences in rates of 30- or 90-day all-cause mortality, bacteremia recurrence, length of hospital stay, readmission, infection complications, adverse events, or emergence of antibiotic resistance, based on systematic reviews and meta-analyses. A subgroup analysis was unable to identify specific patient groups that would benefit from a longer course of therapy.

A 2025 systematic review and meta-analysis of four RCTs evaluated whether a 7-day course of antibiotics was non-inferior to a 14-day course for gram-negative bloodstream infections. The primary outcome was 90-day all-cause mortality. The meta-analysis included an intention-to-treat analysis (n = 3,729) and a per-protocol analysis (n = 3,126). In the intention-to-treat analysis, the median patient age range was 67 to 79 years and 51% of the patients were female. Most patients in this meta-analysis were enrolled in the BALANCE (Bacteremia Antibiotic Length Actually Needed for Clinical Effectiveness) trial, a 2024 RCT that included 2,547 patients with gram-negative bloodstream infections; the remaining patients were from three earlier RCTs. The RCTs had different exclusion criteria, but commonalities included complicated infections (eg, prolonged antibiotic course, uncontrolled infections, polymicrobial infections) and severe immunocompromise.¹

The intention-to-treat and per-protocol analyses used Bayesian statistics to provide a more accurate estimate of the probability of noninferiority. Each result was reported as a risk ratio (RR) with a credible interval. In comparing 90-day mortality between the two groups, 7 days of antibiotic therapy were shown to be noninferior to 14 days in the intention-to-treat analysis (RR = 0.91; 95% credible interval = 0.69–1.22; probability of noninferiority = 97.8%) and per-protocol analysis (RR = 0.93; 95% credible interval = 0.68–1.32; probability of noninferiority = 95.1%).¹

A 2022 systematic review and meta-analysis was published before the BALANCE trial and included the three previously mentioned RCTs (N = 1,186). In addition to the primary outcome of 90-day all-cause mortality, this meta-analysis reported on multiple secondary outcomes, including rates of 30-day all-cause mortality, bacteremia recurrence, length of hospital stay, readmission, infection complications, adverse events, and emergence of antibiotic resistance.²

No significant differences in 90-day mortality or any secondary outcomes were found between patients receiving 7 or 14 days of antibiotics. An analysis of outcomes for several subgroups (eg, males vs females, urinary vs nonurinary source of infection, presence vs absence of hypotension at admission, immunocompromise vs no immunocompromise, age younger than 65 or 65 years and older) showed no significant differences in outcomes for 7 vs 14 days of therapy. This review excluded trials that addressed sources of infection that required prolonged treatment (eg, endocarditis, necrotizing fasciitis, osteomyelitis, abdominal abscess, central nervous system infections, empyema, uncontrolled source of infection).²

A 2025 post-hoc analysis of the BALANCE trial attempted to use clinical decision rules to identify subgroups of patients who might benefit from a longer antibiotic course (7 days or longer). These rules used clinical and complicating factors (eg, immunosuppression, unknown infection source, persistent bacteremia) to develop multifactorial decision tools. These tools were helpful in determining which patients were at increased risk of mortality, but they did not identify groups of patients who had higher 90-day mortality with 7 days of antibiotics compared with 14 days. The results showed that even in patients with a higher risk of mortality, 7 days of antibiotic therapy were sufficient. Because this was a post-hoc analysis, the investigators were limited to the outcomes used in the BALANCE trial and were unable to assess other outcomes (eg, 30-day all-cause mortality, infection recurrence).³