Hospitalized patients with bloodstream infections assigned to 7 days of adequate antibiotic treatment had 90-day mortality that met the prespecified criterion for noninferiority compared with patients assigned to 14 days, according to results from the BALANCE randomized clinical trial.
The investigator-initiated, open-label, randomized controlled noninferiority trial included 3,608 patients at 74 hospitals in 7 countries. Patients were randomly assigned to receive 7 or 14 days of adequate antibiotic treatment, defined as therapy to which the organism was susceptible according to local laboratory reports. Treating clinicians selected the antibiotic agent, dose, frequency, and route.
The primary outcome was death from any cause by 90 days following diagnosis of bloodstream infection, defined as the date of the positive index blood culture. The trial used a 4-percentage-point noninferiority margin, which was smaller than margins used in several prior bloodstream infection trials.
Patients were enrolled from October 2014 to May 2023. The median age was 70 years, 53% were men, and 55% were enrolled in an intensive care unit (ICU). Most infections were community-onset. The most common sources were the urinary tract, intraabdominal or hepatobiliary system, lung, vascular catheter, and skin or soft tissue. Gram-negative monomicrobial infections accounted for 71% of cases, and Escherichia coli was the most common pathogen.
By 90 days, 15% of patients assigned to 7 days of treatment had died compared with 16% of those assigned to 14 days. Per-protocol and modified intention-to-treat analyses were consistent with the primary analysis. In the per-protocol analysis, mortality was 13% in the 7-day group and 15% in the 14-day group. The modified intention-to-treat analysis excluded patients who died prior to day 7, before differences in assigned treatment duration could meaningfully emerge.
The findings extend prior evidence supporting shorter antibiotic courses into critically ill populations, since more than half of enrolled patients were in the ICU at the time of diagnosis. Earlier randomized trials in this area largely excluded ICU patients or enrolled relatively few critically ill patients.
Secondary outcomes were generally similar between groups, including in-hospital death, ICU death, bacteremia relapse, antimicrobial-related adverse outcomes, Clostridioides difficile infection, and secondary infection or colonization with antibiotic-resistant organisms. Patients assigned to 7 days had a median of 19 antibiotic-free days by day 28 compared with 14 days among patients assigned to 14 days of therapy. The researchers noted that secondary outcome analyses were exploratory and were not adjusted for multiplicity.
Prespecified subgroup analyses were generally consistent across enrollment location, acquisition setting, pathogen type, illness severity, vasopressor or inotrope use, frailty score, and source of bacteremia. However, confidence intervals were wide in several smaller subgroups, particularly among gram-positive infections, and the trial was underpowered to determine whether longer treatment may benefit selected patient populations.
The findings should not be applied to patients or infections excluded from the trial, including those with severe immunocompromise, prosthetic heart valves, synthetic endovascular grafts, endocarditis, osteomyelitis, septic arthritis, undrained abscesses, retained prosthetic-associated infections, fungemia, Staphylococcus aureus bacteremia, or Staphylococcus lugdunensis bacteremia. The researchers noted that S aureus bloodstream infections were excluded because of the organism's propensity for metastatic infection and observational data suggesting higher relapse risk with shorter treatment durations.
The trial also had limitations, including its open-label design and protocol nonadherence. Antibiotics were continued longer than assigned in 23.1% of patients in the 7-day group and 10.7% of patients in the 14-day group. Overall nonadherence occurred in 23.9% and 16.5% of patients, respectively. The median treatment duration in the 7-day group was 8 days rather than 7 days, reflecting the pragmatic nature of the trial. The researchers also noted that late C difficile infections may have been missed and that antimicrobial-resistant organisms were assessed through routine specimens rather than active surveillance.
The researchers said the findings support antibiotic stewardship efforts because adopting a shorter treatment strategy would not require new medications or technologies and could reduce antimicrobial exposure and drug-acquisition costs.
"A noninferiority trial can never prove that outcomes are identical in the two groups," the researchers wrote. They added that further research is needed to test individualized and potentially shorter durations so each patient receives just as long a course as is needed.
The study was funded by the Canadian Institutes of Health Research, Physicians Services, the Ontario Ministry of Health Alternate Funding Plan Innovation Fund, the Canadian Frailty Network, the Australian National Health Medical Research Council, and the New Zealand Health Research Council. Full disclosures are available with the study.