Issue: February 2019
February 22, 2019
5 min read

Should physicians prescribe antivirals to patients at high risk for flu-related complications more than 2 days after symptom onset?

Issue: February 2019
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Click here to read the Cover Story, “Bad flu seasons test US hospitals”.


Antivirals should be prescribed to high-risk patients even if they have been sick for more than 2 days.

The release of the latest iteration of the Infectious Diseases Society of America guidelines on the management of influenza are a welcome addition that clinicians can employ to optimally diagnose and treat patients with influenza, an infectious disease that kills thousands of Americans annually. Of the myriad topics covered in the guidelines, one of the most important is antiviral treatments.

Specific antiviral treatments for influenza have existed since the 1960s, with the approval of the first adamantane class antiviral, but recent decades have shown increasing use and familiarity with these drugs as more effective, safer classes of neuraminidase inhibitors (NAIs) and polymerase inhibitors were developed.

An important fact to remember is that randomized trials with NAIs, which were mostly performed before the 2009 pandemic, were placebo controlled in healthy outpatients with uncomplicated influenza. These trials largely showed a 1- to 2-day decrease in duration of symptoms when given within a 48-hour window of symptoms. This 48-hour window, at the time of the pandemic and even today, is a medical dictum that has been seared into the minds of physicians.

Amesh A. Adalja, MD
Amesh A. Adalja

Although the 48-hour parameter is often remembered, the type of patients included in the randomized trials is not. Healthy outpatients with uncomplicated influenza are not equivalent to hospitalized, pregnant, immunosuppressed or medically complex patients who may have pneumonia, respiratory distress or respiratory failure. Applying a 48-hour treatment window in a disparate group of patients to a whole other group of patients is not warranted and not logically defensible.

Randomized trials in at-risk populations in which a placebo is used would be ethically difficult, so, in the modern era, observational studies are where new data in these populations will arise. These data, although less valued than randomized control data, are not devoid of value. In fact, these data illustrate that in high-risk patients such as pregnant women, antiviral therapy can be lifesaving. Although it is true that more prompt antiviral therapy is preferred — as is the case with the treatment of most infectious diseases — later treatment is not without benefit.

Similar to the NAIs, the polymerase inhibitor baloxavir was initially tested in uncomplicated patients and produced similar benefits. There are plans, however, to test the drug in high-risk populations, and these data will be instrumental in shaping views regarding the benefits of antiviral therapy.

The guidelines appropriately recommend treatment in hospitalized patients, patients with severe or progressive illness and patients with chronic medical conditions; immunocompromised patients; pregnant women (through 2 weeks postpartum); children aged younger than 2 years; and adults aged 65 years and older, irrespective of illness duration.

Influenza is a disease that poses the greatest pandemic risk, and as last season demonstrated, the health care industry is not yet adept at managing it. Antiviral prescribing will be a pillar of pandemic planning, and the first step at being able to use these tools in a pandemic is being able to use them appropriately on a daily basis.

Amesh A. Adalja, MD, is an IDSA member and senior scholar at the Johns Hopkins Center for Health Security. His Twitter handle is @AmeshAA. Disclosure: Adalja reports no relevant financial disclosures.

Editor's note: Baloxavir marboxil has been tested in patients at higher risk for influenza-related complications. The results of the study were presented at IDWeek 2018.


Ison MG, et al. Abstract LB16. Presented at: IDWeek; Oct. 3-7, 2018; San Francisco.



The new IDSA guidelines raise more questions than they answer about antiviral treatment.

Clinical practice guidelines are important documents, and their recommendations often have a deep reach within the medical specialty they serve. But sometimes they can raise more questions than they answer, which is what happened to me before I had finished the very first page of the new IDSA guidelines for influenza.

The guidelines note: “Early treatment with antivirals reduces the duration of symptoms and risk of some complications (bronchitis, otitis media, and pneumonia) and hospitalization, and may decrease mortality among high-risk populations.” There are four references provided to support this statement; three come to the opposite conclusion and the fourth is at odds with prior studies. The first, a 2012 study by Hsu and colleagues, concluded that the overall body of evidence about the effect of oseltamivir on mortality is “limited by risk for confounding and selection, reporting, and publication bias.” Moreover, “the confidence in the estimates of the effects [on mortality] for decision-making about using NAIs is low to very low.” That does not seem supportive.

Jeremy Brown, MD
Jeremy Brown

The second, a 2014 Cochrane review led by Jefferson, could not find “any credible evidence that either oseltamivir or zanamivir reduced the risk of complications of influenza, particularly pneumonia, nor reduce risk of hospitalization or death.” Then comes a 2013 study of hospitalized patients by Muthuri and colleagues, and things start to get really weird. It found no significant difference in mortality when comparing any NAI treatment with none. Yet when comparing early and late NAI treatment, it found a significant (and rather impressive) reduction in mortality “albeit with significant publication bias.” It would be challenging to find a plausible biological mechanism to explain these differences, so the authors suggest it was a result of “confounding related to treatment propensity.” This surely questions the analysis in the first place. And finally, a 2014 meta-analysis by Muthuri and colleagues reported a very different result: When comparing any NAI treatment with none, there was a significant decrease in mortality (though not in children, and not when the NAI was started more than 2 days after the onset of symptoms).

That is not all. The most common scenario facing an emergency physician or primary care provider is not whether to treat a severely ill influenza patient with NAIs. Rather, it is whether to start at all in an otherwise healthy patient. In 2009, the IDSA suggested considering NAIs, and the strength of this recommendation was A-I. When facing the same clinical question over a decade later, the advice was the same, but the strength of the recommendation had decreased to C-I; that is a drop from good evidence to poor evidence. So, we are more ignorant about the role for NAIs than we were 10 years ago.

When treating the very sickest influenza patients, providers will continue to use NAIs because, well, maybe they can help. That is certainly not the case for the otherwise heathy patient in whom they should not be prescribed. Even when — and I know this is difficult — they ask for them by name.

Jeremy Brown, MD, is an emergency physician and director of the Office of Emergency Care Research at the NIH. He is the author of Influenza: The Hundred-Year Hunt to Cure the Deadliest Disease in History. Brown can be reached at Disclosure: Brown reports no relevant financial disclosures. Disclaimer: The views expressed by Brown are his own and do not represent those of the federal government.