JOURNAL ARTICLE EDITOR'S CHOICE

Another Point Goes to Team Short Course: Implications of the BALANCE Trial Post Hoc Analysis for Pseudomonas aeruginosa Bloodstream Infections

Leila S Hojat, Emily S Spivak

Clinical Infectious Diseases, Volume 83, Issue 1, 15 July 2026, Pages 48–50, https://doi.org/10.1093/cid/ciag145

Pseudomonas aeruginosa (PA) is one of the most important contributors to the morbidity and mortality attributable to multidrug-resistant infections globally. Tension often arises between concern for undertreatment of PA and overtreatment leading to adverse events and increasing resistance. In the absence of data supporting the alternative, extended treatment duration for PA is often given precedence.

The presupposition that PA requires prolonged treatment is pervasive throughout infectious diseases guidelines. In 2019, it was established that most cases of community-acquired pneumonia need only be treated for 5 days, though a 7-day course was recommended for PA in alignment with the then recently published ventilator-associated pneumonia guidelines [1]. The iDIAPASON study raised further questions of whether a shorter 8-day course for ventilator-associated pneumonia secondary to PA could truly be considered noninferior to a 15-day course [2]. Regarding management of bloodstream infections (BSIs), similar concerns were discussed in the first consensus guidance developed for uncomplicated gram-negative bacteremia, in which 7 days of treatment was recommended for all gram-negative organisms, but the lack of data for non-lactose-fermenting bacteria such as PA was acknowledged, and furthermore it was suggested that these pathogens may be less likely to meet the criteria for uncomplicated BSI [3]. Up to that point, just 1 randomized controlled trial comparing 7 to 14 days of therapy for gram-negative BSI had included PA infections, and only 48 patients with PA were ultimately enrolled [4].

The BALANCE trial results, published in October 2024, supported a 7-day course of treatment for BSI for nearly all bacteria with few exceptions for immunocompromised patients, presence of endovascular material, and infectious syndromes and pathogens for which prolonged treatment is generally recommended [5]. The study enrolled 157 patients with PA, which is the largest population to date of PA infections included in a randomized controlled trial supporting shorter duration therapy. The study investigators performed this post hoc analysis to determine if application of the overall trial results to PA was reasonable, given it represented only 4.5% of the full BALANCE cohort.

Consideration of the PA subset relative to the entire BALANCE cohort illuminates several noteworthy characteristics. Hospital-onset BSI accounted for 53% of PA infections, more than twice that of all BSI cases. The PA cohort had higher rates of pneumonia (22% vs 13%) and more patients requiring mechanical ventilation at enrollment (27% vs 21%), while the full cohort had higher rates of urinary tract infection (42% in full cohort vs 27%) and more patients who underwent a source control procedure (45% vs 39%). Groups had similar Sequential Organ Failure Assessment (SOFA) scores and intensive care unit status at enrollment, as well as similar age, comorbidities, and frailty. Ultimately, the primary outcome of 90-day mortality was higher for patients with PA BSI, occurring in 25% versus 15% among all BALANCE cases, as was the rate of ineffective empiric antibiotics (25% vs 10%).

The authors additionally provided information on the PA subset characteristics after stratification into 7- and 14-day groups. Importantly, randomization did not specifically stratify for PA, and as a result several major differences were evident between groups. Certain factors suggested higher complexity in the 7-day arm, in particular a higher rate of hospital-onset BSI (61% vs 46%). Other notable differences included immunocompromised status, intensive care and mechanical ventilation at enrollment, source control procedures performed, and comorbidities such as diabetes, asthma, and alcohol use, all favoring the longer-duration group.

The primary outcome of 90-day all-cause mortality occurred in 22 patients (29.7%) in the 7-day group versus 17 (20.5%) in the 14-day group (95% confidence interval [CI], −4.8 to 23.2). Other outcomes occurred infrequently with only 2 patients having relapsed BSI, 10 patients developing multidrug resistance, and 5 patients with Clostridioides difficile. The only meaningful statistical difference detected between the 7- and 14-day treatment arms was 17 versus 13 antibiotic-free days among those alive by day 28, respectively (95% CI, 1.6–6.4). The multivariable regression model likewise did not indicate higher odds of mortality based on antibiotic duration (odds ratio, 1.47 [95% CI, .66–3.27]), even among patients with a higher quick SOFA (qSOFA) score. The authors concluded that a 7-day course of treatment resulted in similar outcomes for PA BSI, though they acknowledged the limitation of a post hoc analysis with a small sample size underpowered to detect differences between subgroups.

Interpretation of these findings first requires additional scrutiny of the choice of 90-day all-cause mortality as the BALANCE trial primary outcome. While 90-day mortality is straightforward as an endpoint, it may not be the most relevant for PA BSI, as these patients are likely to have complex underlying issues that may predispose them to additional complications and non-infection-related mortality. Other more clinically relevant outcomes measured, including recurrent bacteremia, occurred infrequently in this subset, suggesting that proximal infection-related outcomes are uncommon overall in PA BSI, and outcomes such as recurrent primary site infection or distant suppurative complications were not feasible to collect in this type of study.

A notable limitation of this analysis is the exclusion of severe neutropenia and transplant populations from the BALANCE trial. Immunocompromised patients are at higher risk for PA and have been observed to receive prolonged PA treatment courses, with a median duration of 15 days in a large multicenter study [6]. Despite excluding 2 important subsets of the immunocompromised patient population, to conclude that the BALANCE trial offered no data for immunocompromised patients would be inaccurate, as 15% of patients received chemotherapy or high-dose glucocorticoids in the PA subset and 12% in the full cohort. Also, other relative immunodeficiencies were prevalent, including diabetes, malignancy, chronic kidney disease, and chronic liver disease. While dedicated investigations are needed to explore outcomes associated with treatment durations in this population, their modest representation in the BALANCE cohort suggests that shorter durations for BSI in immunocompromised individuals are likely justified.

Bearing in mind the limitations, how should clinicians interpret the findings of this analysis? The small sample size restrains the certainty of ruling out a more subtle difference between groups; however, with a baseline affected proportion of 20% in the 14-day PA group, an approximately 20% higher mortality difference would have been detected in the 7-day group, which could conceivably have occurred for patients with inadequately treated systemic PA infection, who are known to have exceptionally high mortality rates. The outcomes were similar also despite the incomplete randomization, which distributed more disadvantageous characteristics into the 7-day arm. Moreover, the 20%–30% mortality rate observed in this study is compatible with mortality rates of patients with PA BSI described in the literature. Therefore, the PA subset demonstrated outcomes generally as expected irrespective of treatment duration.

These findings must be considered in context of the full body of evidence regarding PA BSI–associated outcomes based on treatment duration. A recent systematic review and meta-analysis of 5 retrospective studies pooled data for nearly 1800 patients with PA BSI [7]. The meta-analysis revealed no difference in the propensity-weighted composite outcome of 30-day mortality or recurrent infection between short and long treatment duration groups. Additional clinical outcomes that were more favorable for the shorter duration group included shorter length of stay and fewer adverse antibiotic effects. In one of the included studies, subsequent acquisition of highly resistant PA within 180 days was also numerically higher in the prolonged-course group [8]. A statistically higher odds of new resistance was not identified given the small sample size, though prior studies have demonstrated development of PA resistance with a dose-dependent increase in antipseudomonal treatment [9].

Collectively, the results of the BALANCE trial post hoc analysis of BSI secondary to PA should be considered supportive of shorter treatment durations when positioned among a growing body of literature consistently showing similar outcomes regardless of treatment duration. These data are furthermore emerging in parallel with a gradual shift in prescribing norms toward treatment based on clinical response, with diminishing significance of detection of a pathogen in blood as a meaningful clinical characteristic in itself. In terms of modifiable factors, rapid initiation of active therapy and achievement of source control are likely to affect outcomes for PA BSI, whereas prolonging antipseudomonal therapy only serves to produce more adverse events and perpetuate antimicrobial resistance. Thus, we should embrace shorter treatment for PA BSI where there are no specific factors explicitly suggesting that a short course might lead to failure.

The BALANCE trial investigators have provided substantial contributions to the stewardship literature, and this analysis enhances our ability to apply evidence-based practices to PA BSI, moving past dogmatic treatment paradigms. Yet, despite the encouraging findings of this analysis, larger trials are needed to support definitive conclusions. The SHORTEN-2 trial was designed to address this gap by comparing 7 versus 14 days of therapy among patients with PA BSI [10]. The study has enrolled its target of 306 patients and completed the follow-up period, with results anticipated later in 2026. Meanwhile, antimicrobial stewards must rise to the challenge of assisting clinicians with understanding the implications of these results. Acceptance of shorter courses of antimicrobial therapy will require a fundamental shift away from fixed durations for BSI, toward treatment guided by the underlying infectious syndrome and clinical response.