Acute respiratory illness is a common cause of hospitalization in adults and a common reason for antibiotic prescriptions in hospitalized patients, despite the fact that many respiratory conditions are caused by viruses.
Diagnostic uncertainty is a big part of the problem. With traditional testing for respiratory viruses taking 24 to 48 hours, physicians faced with a sick patient often resort to antibiotic therapy even though it may not be needed. As a result, treatment for acute respiratory conditions is considered a ripe target for antibiotic stewardship, one that could benefit from diagnostic tools that can quickly determine whether the cause of illness is bacterial or viral.
The theory is that if physicians can find out in real time that a patient with acute respiratory symptoms has a viral infection, they'll be much less likely to prescribe antibiotics. To date, however, there has been little evidence to back this theory.
Secondary outcomes show promise
So over the winters of 2015 and 2016, a team of researchers from the University of Southampton in England enrolled 720 adult patients at Southampton General Hospital in a randomized controlled superiority trial to test this theory.
The patients were randomly assigned to two groups, with 362 receiving a rapid point-of-care test and 358 receiving standard clinical care for acute respiratory illness. The researchers used a molecular test called the FilmArray Respiratory Panel, which processes swabs for 17 viruses (including influenza A and B) and three types of bacteria that cause upper respiratory tract infections. The test can produce a result within an hour.
The primary outcome of the study was the proportion of patients treated with antibiotics during hospitalization. Secondary outcomes included duration of antibiotic use and the proportion of patients who received only a single dose of antibiotics or were treated with antibiotics for less than 48 hours. The researchers also looked at other potential clinical benefits of the test, including hospital length of stay and the proportion of confirmed influenza patients who received appropriate antiviral treatment (neuraminidase inhibitors like oseltamivir [Tamiflu]).
The results, published yesterday in The Lancet Respiratory Medicine, did not quite support the hypothesis, at least in terms of the primary outcome. Of the patients in the testing group, 84% received antibiotics, compared with 83% of the patients who received standard care. In addition, the mean duration of antibiotic use didn't differ between the two groups.
But those results didn't necessarily surprise corresponding author Tristan Clark, MD, an associate professor and honorary consultant in infectious diseases at the University of Southampton, because some of the patients who received point-of care testing were given antibiotics before the results came back. Some, in fact, had received antibiotics before randomization occurred.
"This is a consequence of the current pathways for hospitalized patients with ARI (acute respiratory infection), where antibiotics are started almost as soon as patients come through the door," Clark told CIDRAP News.
But when Clark and his colleagues looked at secondary outcomes and dug deeper into the numbers, the results were more promising. For example, a greater proportion of patients in the testing group received a single dose of antibiotics compared with the control group (10% vs 3%), and more patients in the testing group received a short course of antibiotics (17% vs 9%).
Clark said this finding didn't translate into a statistically significant overall reduction in duration of antibiotics because it was offset by prolonged antibiotic use in certain patient groups, such as those with pneumonia, which was not altered by point-of-care testing. But in patients with asthma and chronic obstructive pulmonary disease who were tested, antibiotic use was reduced by 1 to 2 days. "No other stewardship trials have been able to achieve this, to my knowledge," said Clark.
Additional analysis of the patients who didn't receive antibiotics before randomization or before receiving the test results showed that only 51% of those patients were treated with antibiotics, compared with 64% in the control group. Patients who tested positive for a viral infection also received shorter courses of antibiotics.
Compared with standard care, rapid point-of-care testing was also associated with a higher detection rate of viruses, reduced length of stay, quicker detection of the flu, and more appropriate antiviral treatment for influenza-positive patients. The study suggests that correct treatment with neuraminidase inhibitors was more common among influenza-positive patients in the testing group, with those patients getting the first dose of the drugs more quickly than those in standard care.
Can diagnostics change practice?
Clark said that although the primary outcome was negative, he's delighted by the findings, which indicate to him that if hospitals can do point-of-care testing on patients as soon as they arrive at the hospital, they could reduce inappropriate antibiotic use and improve care. "It shows what I always suspected—that knowing what infection your patient has right at the start of their inpatient stay improves their outcome and is better for the hospital," he said.
Clark also suggested that ultimately, early cessation of antibiotic treatment may be the better outcome to measure, since withholding antibiotics from patients until test results arrive could potentially worsen outcomes.
Andrew Pavia, MD, a professor of pediatric infectious diseases at the University of Utah and author of an accompanying commentary, says that while the results are encouraging, further proof is needed that improved diagnostic tools can help reduce antibiotic use in acute respiratory infections.
"We must prove that improved diagnostics change practice," Pavia writes. "We must also show economic and clinical value. Until then, clinicians might continue to entertain the offered fallacy."
Clark and his colleagues note that the findings need to be replicated in larger, multicenter studies before any definitive conclusions can be made.
Apr 5 Lancet Respir Med abstract
Apr 4 Lancet Respir Med commentary