This systematic review and meta-analysis aimed to determine whether a 7-day course of antibiotics is noninferior to a 14-day course for treating Gram-negative bacteremia secondary to UTIs.

McAteer et al. studied 1099 hospitalized adults with complicated UTIs and bacteremia. Recurrence of UTI rates within 30 days were similar for 10 versus 14 days of therapy (adjusted odds ratio [aOR] 0.99, 95% CI 0.52–1.87). As for therapy, the subgroup explicitly counted intravenous (IV) β-lactam for the entire course or for a step-down to highly bioavailable oral agents, defined as: ciprofloxacin, levofloxacin, trimethoprim–sulfamethoxazole, amoxicillin, amoxicillin–clavulanate, and cephalexin. However, 7-day treatment was associated with higher recurrence (aOR 2.54, 95% CI 1.40–4.60), unless patients received highly bioavailable agents, in which case outcomes were comparable (aOR 0.76, 95% CI 0.38–1.52) [7]. Tingsgård et al. conducted a target trial emulation in 1040 patients, 78% of whom had UTIs. Comparing short-course (5–7 days) with long-course (8–14 days) therapy. Therapy defined as “adequate empiric” IV regimens listed as acceptable were the following: piperacillin–tazobactam, ampicillin + gentamicin, mecillinam, cefuroxime, ceftazidime, cefotaxime, meropenem, ertapenem, tobramycin, and metronidazole. They found similar 90-day mortality (11.3% versus 10.0%; risk difference 1.3%, 95% CI −0.7 to 3.3) and relapse rates (0.7%, 95% CI −2.3 to 3.8) [12]. Von Dach et al. randomized 504 adults with uncomplicated Gram-negative bacteremia (mostly UTI source) to 7 or 14 days of therapy. They followed local guidelines for antibiotic choice and administration route: piperacillin–tazobactam and amoxicillin–clavulanate, cephalosporin and fluoroquinolone (appeared in bivariable analysis of empirical therapy), and IV to oral was allowed per routine practice. The 7-day regimen was noninferior, with 30-day clinical failure rates of 6.6% versus 5.5% (difference 1.1%, one-sided 97.5% CI −∞ to 6.3), and similar 90-day failure rates (10.6% versus 10.5%) [17]. Yahav et al. randomized 604 patients (68% UTI source) and found that the 90-day composite outcome occurred in 45.8% (7 days) versus 48.3% (14 days) (risk difference −2.6%, 95% CI −10.5 to 5.3), again confirming noninferiority [10]. Choice of empirical/directed agent was left to the treating physicians; IV to oral switch timing was also at the discretion of the physician (no protocol-mandated drugs). Highly absorbable antibiotics such as quinolones were used. Chotiprasitsakul et al. analyzed 248 matched patients with Gram-negative bacteremia, mostly UTI-related, and found no significant difference in 90-day mortality (11% versus 12%) or relapse (3% versus 4%) between 7- and 14-day therapy [11]. Specific agents/routes were not enumerated, and aminoglycoside monotherapy was excluded. There was no specification on administration route. The BALANCE trial, the largest to date with 3608 patients (42% UTI source, 67% Gram-negative), found 90-day mortality of 14.5% (7 days) versus 16.1% (14 days) (difference of −1.6 percentage points, 95.7% CI −4.0 to 0.8), confirming noninferiority across key subgroups [4]. Selection of antibiotic, dose, frequency, and route was at the discretion of the clinician (no mandated agents; route not fixed), and any route was permitted per physician’s choice.

While the overall findings of this meta-analysis are consistent with the conclusions of most individual studies—supporting the noninferiority of 7-day antibiotic therapy for Gram-negative bacteremia from a urinary source—some discrepancies merit attention. Notably, although several randomized trials such as Yahav et al., von Dach et al., and the BALANCE trial demonstrated noninferiority across both 30- and 90-day outcomes, our meta-analysis did not confirm noninferiority for 30-day recurrence of bacteremia or mortality. For example, the pooled RR for 30-day recurrence was 1.38 (95% CI 0.80–2.37), exceeding the prespecified noninferiority margin of 1.25 and raising the possibility that short-course therapy may be associated with a higher recurrence risk in the early post-treatment period.

This divergence may stem from heterogeneity in population characteristics, clinical severity, antibiotic regimens, and definitions of recurrence across studies. Additionally, the inclusion of observational cohorts in the pooled analysis—particularly McAteer et al., which identified increased recurrence of UTI with 7-day therapy in certain subgroups—may have introduced residual confounding that is less prevalent in controlled trial settings. Observational studies also varied in antibiotic selection and bioavailability, which could influence recurrence rates independently of duration.

Further, while the 90-day recurrence analysis approached statistical significance in favor of short-course therapy (RR 0.68, 95% CI 0.45–1.01; p = 0.06), this signal was not uniformly observed across all trials. Interestingly, in sensitivity analyses limited to observational data (e.g., Tingsgård et al.), the 7-day course appeared superior to 14 days for preventing 90-day recurrence. However, this finding must be interpreted cautiously owing to the potential for selection bias, unmeasured confounders, and differences in analytical methodology, including target trial emulation with cloning and censoring.

This meta-analysis supports the use of 7-day antibiotic therapy for stable patients with Gram-negative bacteremia from a urinary source. The findings align with current expert recommendations and highlight the potential to reduce treatment duration without compromising clinical outcomes. Adopting shorter courses may help limit antibiotic exposure, reduce adverse effects, and support antimicrobial stewardship efforts, especially in E. coli bacteremia, which accounted for the majority of cases.

Short-course therapy may also reduce hospital stays and resource use. However, caution is warranted in generalizing these findings to patients with non-urinary sources, immunocompromised hosts, or those without adequate source control. The limited representation of pathogens such as Pseudomonas and Enterobacter species restricts broad microbiologic applicability.

Half of the included studies were observational, introducing possible bias and confounding. While sensitivity analyses attempted to address this, variability in study design and treatment regimens may have influenced outcomes. Additionally, 90-day follow-up may not accurately reflect the impact of antibiotic duration, as most relapses occur within 30 days. Regional resistance patterns, which were not consistently reported, may also affect optimal treatment strategies.

In summary, 7-day antibiotic therapy appears sufficient for most cases of urinary-source Gram-negative bacteremia, but clinical judgment remains essential. Future trials should evaluate duration by pathogen, severity, and host status to better guide individualized care.