Algorithm may help guide antibiotic therapy for staphylococcal bacteremia

The algorithm predefined diagnostic evaluations, antibiotic selection, and duration of therapy, and its outcomes were compared to those in similar patients whose physicians had unrestricted choice of antibiotic and duration of therapy.

An algorithm may help guide antibiotic treatment for patients with staphylococcal bacteremia and decrease length of therapy without increasing rates of serious adverse events, a new study found.

Researchers conducted a randomized trial from April 2011 to March 2017 that included 509 patients (mean age, 56.6 years; 44.4% women) at 15 academic medical centers in the U.S. and one in Spain. Eligible patients were ages 18 years and older and had one or more positive blood cultures for Staphylococcus aureus or coagulase-negative staphylococci. Patients who had known or suspected complicated infection at the time of randomization were excluded.

Participants were randomized to receive algorithm-based therapy (n=255) or usual care (n=254). The algorithm group had predefined diagnostic evaluations, antibiotic selection, and duration of therapy, whereas clinicians treating those in the usual practice group had unrestricted choice of antibiotics, duration, and other aspects of clinical care. Participants with S. aureus were followed for 42 days after the end of therapy, and those with coagulase-negative staphylococcal bacteremia were followed for 28 days.

The primary outcomes were clinical success, as determined by a blinded adjudication committee (tested for noninferiority within a 15% margin), and serious adverse event rates in the intention-to-treat population (tested for superiority). The secondary outcome measure (tested for superiority) was the number of antibiotic days among per-protocol patients with simple or uncomplicated bacteremia. Results were published online on Sept. 25 by JAMA.

Overall, 480 (94.3%) participants completed the trial. Clinical success was documented in 82.0% of patients assigned to receive algorithm-based therapy and 81.5% of those receiving usual care (difference, 0.5%; one-sided 97.5% CI, −6.2% to ∞), meeting the criteria for noninferiority. Serious adverse events were reported in 32.5% of algorithm patients and 28.3% of usual practice patients (difference, 4.2%; 95% CI, −3.8% to 12.2%). For the secondary outcome, mean duration of therapy was 4.4 days for algorithm-based therapy compared to 6.2 days for usual practice (difference, −1.8 days; 95% CI, −3.1 to −0.6). Limitations of the study include its open-label design and the possibility that repeated exposure to the algorithm influenced clinicians' subsequent management decisions in both groups of patients, the study authors noted.

These results will likely influence future treatment guidelines, according to an accompanying editorial. “However, algorithms cannot simply be applied in a vacuum without ongoing monitoring and adjustment based on an individual patient's clinical course. … [F]uture investment in clinical trials targeting optimal antibiotic selection and duration are essential to continued progress,” the editorialists wrote.