Can bacteremic patients be treated for 7 days rather than 14?
BALANCE trial and three smaller RCTs provide guidance
More than half a million bloodstream infections occur annually in the U.S. and are estimated to cause nearly 100,000 deaths, with worldwide totals many times higher.
Treatment courses for many severe infections have been reduced in length with no apparent detriment, but most trials testing antibiotic durations have excluded patients with bacteremia. Surveyed infectious disease experts advise about 14 days of antibiotic treatment for most patients with bloodstream infections, especially the critically ill.
Recent randomized trials have sought to discover whether shorter antibiotic courses can result in similar cure rates for patients with bloodstream infections—including the most severely ill.
The BALANCE Trial
At 74 hospitals in 7 countries over 9 years, investigators randomized 3,608 hospitalized patients with confirmed bacteremia to receive either 7 or 14 days of antibiotics, with agents chosen by their treating physicians.
Treatment allocation was blinded until day 7 but unblinded after that, because placebos could not be arranged for all the various antibiotic regimens.
Almost 75% of the patients were infected with gram-negative bacilli, with other bacteria each making up about 2 to 7% of the total infections. Patients with Staphylococcus aureus were excluded, as were the severely immunocompromised and those with non-removable indwelling devices (e.g., endografts).
Just over half of the patients were in the ICU at the time of diagnosis. Urinary tract infections were the most common source, at about 42% of the total.
After 90 days of follow-up, the 7-day course was non-inferior to the 14-day course for the primary endpoint of all-cause mortality (14.5% vs. 16.1%). This met the pre-specified non-inferiority margin of four percentage points or less.
There were no significantly increased rates of C. difficile infection or other adverse events in the 14-day antibiotic arm (2% vs. 1.7% in the 7-day group).
Subgroup analyses did not identify any patient population that might benefit from longer courses of antibiotic therapy, or who were harmed by a shorter course. However, confidence intervals were wide and could not rule out inferiority of the 7-day strategy in patients with vascular catheters, high frailty scores, gram-positive or polymicrobial infections, or very high organ failure scores.
The trial was not able to demonstrate lower antibiotic resistance in patients assigned to the 7-day course.
Clinicians frequently deviated from the treatment protocol, with 23% of patients in the 7-day group and ~11% in the 14-day group receiving longer treatment courses. Thus, the BALANCE trial might be best interpreted as showing that an intended 7-day treatment course is non-inferior to 14 days for bacteremic patients who are clinically improving.
Other Randomized Trials Testing Shorter Antibiotic Courses for Bacteremia
Three recent medium-sized randomized trials also concluded that 7-day courses of antibiotics were non-inferior to 14 days:
Yahav et al Clin Infect Dis 2019 (n=604)
von Dach et al JAMA 2020 (n=504)
Molina et al Clin Microbiol Infect 2022 (n=248)
However, they enrolled few ICU patients, used large noninferiority margins (10%), questionable endpoints (e.g., days of treatment), or clunky composite outcomes (e.g., all-cause mortality/relapse/suppurative or distant complications/readmission/extended hospitalization). These techniques tended to reduce confidence in their conclusions.
The BALANCE authors combined the three aforementioned randomized trials with their own data (which comprised 87% of the weight), and performed a meta-analysis using the single outcome of 90-day mortality. The 7-day strategy was highly likely to be non-inferior to 14 days of therapy in that analysis.
Conclusions
It seems likely that for most patients with gram-negative bacteremia, an intended 7-day treatment course—extended when appropriate—will result in cure rates equivalent to longer intended treatment courses.
The BALANCE trial did not show any clear danger from 7-day antibiotic courses in the smaller number of patients infected with Enterococcus (~7%), Streptococcus (~9%), or other species, but S. aureus bacteremias were excluded. Clinicians often treated patients for longer than the originally intended courses in both arms.
For most serious infections (e.g., pneumonia and pyelonephritis), antimicrobial treatment courses have been shortened over time without apparent harm. The BALANCE trial and its smaller cousins might give more clinicians the confidence to shorten antibiotic courses for patients with bacteremia other than S. aureus who lack indications for longer periods of treatment.
Would be nice to see a sub-hoc analysis measuring procalcitonin levels and its correlation with days of antibiotic treatment. Nice to see we can think about extending this practice to other gram negatives apart from Enterobacteriaceae.