Sepsis study highlights risks of overly broad antibiotic treatment

April 22, 2020

A large new study of patients with sepsis has found that broad-spectrum antibiotics are frequently administered to patients not infected with antibiotic-resistant pathogens and are associated with higher mortality in these patients.

The study, published in JAMA Network Open, looked at data on more than 17,000 culture-positive community-onset sepsis patients in U.S. hospitals and found that more than two-thirds received antibiotics targeting drug-resistant organisms like methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa. Yet only in one in eight sepsis patients had resistant gram-positive or gram-negative organisms, respectively. Treatment with unnecessary broad-spectrum antibiotics was associated with a 22 percent increase in mortality.

The study also found that undertreatment with antibiotics that didn't cover the infecting pathogen was associated with higher mortality.

The authors of the study say the findings highlight the need for more judicious use of broad-spectrum antibiotics in sepsis treatment and for better tests to rapidly identify resistant organisms.

The retrospective cohort study, conducted by researchers with Harvard Medical School, Brigham and Women's Hospital, and the National Institutes of Health, analyzed data on 17,430 adults treated for sepsis at 104 U.S. hospitals from January 2009 through September 2015. The researchers wanted to identify how many of these patients had drug-resistant pathogens in order to determine how often broad-spectrum antibiotics were truly needed, and what the risks of overly broad therapy were in these patients.

Answering these questions is important because sepsis, which occurs when the body's overreaction to an infection causes tissue damage and organ failure, is a leading cause of death in hospitalized patients. Early identification and treatment are critical aspects of effectively managing the condition, and national and international guidelines recommend early empiric broad-spectrum antibiotic therapy as an element of appropriate treatment. The idea behind this strategy is to cover for all likely pathogens.

Yet while sepsis is frequently caused by gram-positive and gram-negative pathogens, it's unclear how many patients who present with sepsis have been infected with the type of resistant bacteria that requires broader antibiotic therapy.

The resistant pathogens assessed in the study included MRSA, vancomycin-resistant Enterococcus (VRE), extended-spectrum beta-lactamase-producing gram-negatives (ESBLs), carbapenem-resistant Enterobacteriaceae (CRE), and ceftriaxone-resistant gram-negatives (CTX-RO). Each potential antibiotic-pathogen combination was assessed using antibiotic susceptibility analyses from in vitro reports generated by each hospital.

The researchers considered antibiotic therapy to be inadequate if at least one pathogen isolated from a patient was not susceptible to all antibiotics. Antibiotics were deemed overly broad if they were active against MRSA, VRE, CTX-RO, and ESBLs, but none of these pathogens were isolated from treated patients.

Of the 17,430 patients included in the analysis, 2,865 (16.4 percent) died in the hospital. The most common pathogens isolated were Escherichia coli (33.7 percent), S. aureus (21.3 percent), and Streptococcus species (13.5 percent). Among 15,183 cases in which all antibiotic-pathogen susceptibility combinations could be calculated, 81.6 percent received adequate antibiotic therapy.

However, while empiric antibiotics (primarily vancomycin and anti-Pseudomonal beta-lactams like ciprofloxacin or levofloxacin) targeted resistant organisms in 67 percent of patients (11,683), resistant organisms were uncommon. The most prevalent resistant pathogens were CTX-RO (13.1 percent, 2,278 patients) and MRSA (11.7 percent, 2,045), followed by VRE (2.1 percent, 360) and ESBLs (0.8 percent, 133). The net prevalence for at least one resistant gram-positive pathogen was 13.6 percent (2,376 patients) and for at least one gram-negative organism was 13.2 percent (2,297 patients).

When the researchers analyzed patient outcomes, adjusting for factors including patient demographics, comorbidities, and microbiologic characteristics, they found that both inadequate empiric antibiotic therapy (odds ratio [OR], 1.19; 95 percent confidence interval [CI], 1.03 to 1.37, P = .02) and unnecessarily broad empiric antibiotics (OR, 1.22; 95 percent CI, 1.06 to 1.40, P = .007) were associated with higher mortality.

The authors suggest the increased mortality in patients who were overtreated with broad-spectrum antibiotics could be linked to the 26 percent increased risk of Clostridioides difficile infection they found in those patients, along with an observed trend toward acute kidney injury, and disruption of the gut microbiome. They also note that the increased mortality was seen only in patients who did not experience septic shock.

They conclude that while it's understandable that guidelines emphasize broad-spectrum antibiotic for sepsis, the findings suggest clinicians need to weigh the risks of unnecessarily broad antibiotic treatment against the risks of inadequate treatment.

"All told, the net fraction of patients with sepsis who would benefit from broad-spectrum therapy, including agents active against both MRSA and Pseudomonas, is small," the authors wrote. "This may be an acceptable trade-off given the increased risk of death associated with inadequate therapy, but it underscores the need for rapid tests to more efficiently identify the small fraction of patients who truly need broad spectrum therapy."

They add that, since a substantial fraction of sepsis patients don't have bacterial infections, further research is needed on unnecessarily broad antibiotic therapy in the culture-negative population.

Visit CIDRAP for more news