By Salynn Boyles

In a subset of patients presenting to emergency departments with suspected sepsis, early administration of antibiotics might have done more harm than good, researchers found.

In an analysis of records for more than 58,000 patients, early antibiotic administration was associated with higher mortality among those with 10% predicted mortality and 20% predicted infection, (absolute mortality difference 1%; 95% CI 0.4%-1.9%), reported Ian Barbash, MD, of the University of Pittsburgh.

With 10% predicted mortality and 60% predicted infection, there was a trend toward lower mortality (absolute difference -0.8%; 95% CI -1.9% to 0.4%).

"Patients with a high degree of certainty regarding infection had better outcomes when treated early with antibiotics, but patients who showed less likelihood of infection seemed to have worse outcomes," Barbash said.

The findings argue in favor of giving critical care clinicians more discretion in determining whether to treat suspected sepsis cases with antibiotics than allowed under current guidelines, he concluded.

Barbash said greater emphasis on the early identification and treatment of sepsis and septic shock through initiatives like the Surviving Sepsis Campaign has greatly benefited patients presenting to emergency departments with severe illness.

But sepsis guidelines, which recommend universal antibiotic administration within specific time frames, may not be as appropriate for patients with less severe symptoms who are suspected of having systemic infection.

On the other hand, he said that while the findings suggest antibiotics may be harming these patients, this is probably not the case.

"It could be that these patients are being treated with antibiotics because they seem sicker to the physician," he said. "Physician worry is not something we can account for in risk adjustment models. These patients may have worse outcomes because they are sicker, but it may have nothing to do with the fact that they received antibiotics."

It is also possible, he said, that the emphasis on treating suspected sepsis in patients who don't have it may delay the diagnosis of the condition that is actually making them sick, which could lead to worse outcomes.

"There is a risk when we develop policies that focus on a specific disease, like sepsis, of taking attention away from other conditions," he said. "The challenge is finding the right balance."

He said early antibiotic administration is usually the right call in patients with well recognized clinical features of infection, such as high fever, high heart rate, and high white blood cell count.

"In patients without these classic clinical features we need to consider the possibility of infection, but also other things," he said, adding that current guidelines may miss important differences in patients suspected of having sepsis.

Barbash and colleagues conducted their analysis in an effort to better understand the impact of antibiotic timing on patients suspected of having sepsis with varying risks for death and infection.

Using electronic health records from 10 academic and community hospitals, they identified 58,413 adults admitted through the Emergency Department with suspected infection within 6 hours of arrival and two or more sequential organ failure assessment points within 24 hours of suspected infection.

Patients were divided into 100 mutually exclusive groups using their deciles of predicted mortality and predicted infection, and the likelihood of antibiotic administration within three hours of suspected infection was determined for each group. A heatmap was then created based on the results.

A second heatmap was created showing the ratio between the median predicted mortality associated with receiving antibiotics versus not receiving antibiotics within 3 hours.

The study included data on 58,413 patients admitted with suspected infection and organ dysfunction. Early antibiotic administration was most associated with the probability of infection, rather than mortality.

When mortality risk was examined, an interaction was shown between early antibiotic administration and the probability of infection.