Influenza in Children: Is Treat Early, Treat Often the Best Strategy?

Weijen Chang, MD, Matthew Garber, MD, Ricardo A. Quinonez, MD

February 25, 2015

Editor's Note: This post originally appeared at the Society of Hospital Medicine's blog, The Hospital Leader.

Weijen Chang, MD: Recently, Thomas Frieden, MD, director, US Centers for Disease Control and Prevention (CDC), in response to criticism about the reduced efficacy of this season's influenza vaccination, has been prominently advocating use of antiviral medications (oseltamivir [Tamiflu®] and zanamivir [Relenza®]) for the treatment of influenza infections in adults and children. In short, his philosophy can be paraphrased as, "Treat early, treat late, treat often." This treatment recommendation, however, seems to be swimming upstream against a growing river of evidence that questions the efficacy of influenza antiviral medications, especially in light of an unfavorable adverse effect profile.^[1]

The current CDC recommendations advocate treatment with antivirals for all children hospitalized with influenza infection, despite recent studies showing lack of efficacy in otherwise healthy children.

Based on my concerns as a pediatric hospitalist with the recommendations and the lack of evidence, I decided to touch base with two other pediatric hospitalists, Matthew Garber, MD, and Ricardo Quinonez, MD, and I asked them about the use of antiviral medications in the treatment of pediatric influenza infections.

Drs Garber and Quinonez, what do you think is driving Dr Frieden's strong, perhaps non-evidence-based, recommendations for antiviral treatments in influenza infection?

Matthew Garber, MD: This is a difficult question, especially because I don't know Dr Frieden and am unaware of his life experiences, general approach to medicine, how he understands and deals with risk, or various outside pressures he may be under.

I can speak to some underlying psychological biases that most of us share and other issues in evidence-based medicine that may be at play. First you have the original CDC,^[2] World Health Organization (WHO),^[3] and American Academy of Pediatrics (AAP)^[4] recommendations that advocate use of antivirals pretty broadly. If you were to take those recommendations at face value, without critically looking at the evidence and the Cochrane review,^[1] you could logically conclude that because this season is predicted to be severe, and the vaccine is not very effective, we need to rely even more heavily on these medicines.

As you know, the best evidence we have—systematic reviews and meta-analyses of randomized controlled trials—including data from methodologically sound, industry-sponsored trials that were withheld from the Cochrane respiratory group for 5 years^[5] and tells us that these medications reduce symptoms by about 1 day if given very early in the course of infection. However, there is no evidence that they prevent complications, hospitalizations, or deaths from influenza. Furthermore, even though prophylaxis with antivirals can prevent symptomatic disease in a contact, prophylaxis has not been shown to decrease transmission of the virus (the main impetus for stockpiling these medicines in order to halt an epidemic).

So did the Cochrane review prove that antivirals do not prevent complications, hospitalizations, and death? That is the evidence-based medicine issue I'm talking about. It is very hard to prove a negative. Except in extreme extenuating circumstances, we generally require proof that a drug works before using it—we do not require proof that a drug does not work to avoid using it.

Then there are the psychological issues to take into consideration. People, and perhaps especially doctors, like to help other people. Doctors have been trained that when a patient is sick, we find out what is wrong with him/her and then find the best treatment and administer it. It is very hard for us to say, "Thank you for coming to see me and paying for this visit, but there is really no effective therapy for this condition other than symptomatic relief." We'd much prefer to say, "Aha! You have the flu, and here is the drug that will make you better." Then of course, especially in pediatrics, the patients do get better, which reinforces our behavior. The placebo effect is large, especially in children,^[6] which also encourages this type of behavior.

Finally, unfairly, we treat errors of omission differently from errors of commission. If someone complains of nausea and vomiting after receiving oseltamivir, we say, "Well yes, that is a known side effect of that drug." Basically it is the cost of doing business. If we don't give an antiviral and the person becomes very sick, we are often faulted for failing to provide the right treatment and may even be sued. Even our language contributes to overtesting and overtreatment saying things like, "To be conservative you better get that chest x-ray. And just to be safe, let's start that antibiotic or antiviral." When in fact one could argue that the conservative path in the face of uncertainty would be to intervene less, not more. This is partly explained by our failure to consider the harms of our interventions (finding an innocent lesion on the chest x-ray that leads to an invasive procedure; Clostridium difficile infection following antibiotics).^[7]

What Is the Role of Professional Societies?

Dr Chang: Given the recent Choosing Wisely® campaign started by the American Board of Internal Medicine (ABIM) and joined by major professional societies (AAP, Society for Hospital Medicine [SHM], etc), should one of the major societies balance the voice of the CDC with their own recommendations?

Ricardo A. Quinonez, MD: The short answer is yes, particularly those societies that use strong evidence-based decisions for their own guidelines. My view is that they should adhere to the highest standard both when they write their own guidelines and endorse others' recommendations. I really think that is not the case currently. To be fair, though, it is going to be extraordinarily difficult for a medical society to contradict decisions made by one of the highest medical authorities in the land, such as the CDC.

Where it really should start is with those medical experts within our government to have a measured response to ongoing health crisis and not swing the pendulum so far that they are ignoring science. I understand they want to seem proactive given the initial criticism the CDC received for handling of the Ebola crisis. However, the failure of this year's vaccine to completely cover the flu represented yet another crisis. The unmeasured response was to support, in an extraordinary way, medications for which evidence is not just lacking but also has a clear track record of manipulation of data by its industry supporters.

I would direct readers to well-documented instances of this behavior with oseltamivir by the British Medical Journal.^[8] I think the CDC and others should take an unbiased approach to recommendations even in times of crisis. The US Preventive Services Task Force (USPSTF) framework for making recommendations is a great example to follow. If evidence is not there to support a particular therapy or test, no recommendation is made. It allows clinicians to use their judgment to make clinical decisions in those instances in which the evidence is equivocal or highly biased, as is the case for antivirals in influenza. The CDC's strong stance on this has really put conscientious physicians in a difficult spot.

Dr Chang: The reality of practicing hospital-based pediatrics is that it is a team effort and, as physicians, we are not immune to the "peer pressure" of others when it comes to treatment selection. How much antiviral use by hospitalists is due to a fear of being second-guessed (especially if the patient subsequently "crashes")?

Dr Quinonez: This is a fantastic point, and I could not agree more with your assessment. I think pediatricians and other physicians have been placed in an extraordinarily difficult situation. Those of us who like to make decisions based on evidence when treating children feel extremely uncomfortable with the CDC's recommendations. On the one hand, the latest meta-analysis of neuraminidase inhibitors not only found insufficient evidence to support an effect on complications in patients with influenza but also found that the harm has been severely understated. These medications may actually put children at risk of developing important neuropsychiatric symptoms and significant emesis.

(In a post-interview exchange, Dr Quinonez added: "A more recent meta-analysis published since this original interview took place suggests increased benefit of oseltamivir compared with the Cochrane investigators' review. It is highly important to note that this most recent review was sponsored by the manufacturers of oseltamivir and conducted by a group of investigators called the Multiparty Group for Advice on Science (MUGAS), and was funded by an unrestricted grant from Roche. This might lead the reader to conclude significant conflict of interest. Of note, the results for benefits in this newer review are presented as relative risks while the results for adverse effects are presented as absolute risk. This, of course, makes the risk-benefit ratio artificially inflated in favor of benefit.")

CDC defends its decisions by pointing to single studies showing decreased complications and mortality but ignores the higher-level evidence (meta-analysis) that should drive recommendations. The US Food and Drug Administration (FDA), an agency that has indeed reviewed the full study data on Tamiflu® during its approval process, has actually contradicted the CDC's recommendations and has forced the makers of oseltamivir to include in its labeling language that effectively states that Tamiflu has not been shown to prevent complications like serious bacterial infections.

On the other hand, not following the CDC recommendation puts practitioners at a very significant risk for legal exposure. It is not hard to imagine that a patient who presents with influenza-like symptoms and is not started on antivirals and later develops a complication (which, again, the evidence does not support antivirals being able to prevent) could result in the physician being dragged into a painful legal proceeding based solely on the strong recommendation from the CDC that seems to be using antivirals as the solution to the inefficacy of this year's flu vaccine.

Putting It Into Practice

Dr Chang: Does the severity of the current influenza season, coupled with the lower-than-expected efficacy of the seasonal vaccine, warrant increased antiviral usage by hospitalists?

Dr Garber: I don't think it does. But to be fair, this argument is not completely without merit. As I discussed before, it's hard to disprove a negative, but with the large number of patients studied, I think we can rule out a large effect size of the antivirals. While antivirals may have a small effect size to prevent complications in a limited group of patients, there is currently no high-quality evidence supporting this. The clinician must weigh several factors when considering interventions for her patients.

Let's pretend, just for the sake of argument, that a particular antiviral was shown to have absolutely no side effects and was free (maybe it's derived from a pesky weed that homeowners would be happy for us to pluck and ingest). Let's also pretend that this year's influenza season is much worse than the 1918 Spanish flu epidemic (imagine Ebola that spreads like the flu). In this imaginary scenario, prescribing the antiviral makes sense. No side effects and it's free. It may help some people avoid complications, but even if it doesn't, what's the harm? Unfortunately, no such medicine exists. They all have side effects, and current recommended antivirals are quite pricey. We know that oseltamivir has a significant rate of emesis and, though rare, occasionally causes severe neuropsychiatric symptoms, even suicide. So those need to be added into the equation, and the more prescriptions we write, the more potential harm we cause.

While harms resonate with doctors and patients more than costs, we also must consider costs. Because the United States does not have unlimited resources, money spent in one area of healthcare will not be available in another. That is the reasoning behind the statement, "Waste is unethical." We need to be good stewards of our country's healthcare dollars and should not allocate large resources to unproven therapies as there will undoubtedly be patients who will be denied effective therapy as a result.

So there is a balancing act. We must weigh the severity of the disease and the evidence supporting the potential of the therapy to improve important outcomes (such as mortality) against potential harms and costs of the therapy. Right now we do not have strong evidence that antivirals improve important outcomes, but they do have significant side effects and high costs. So at this time I don't think the severity of the season coupled with the poor efficacy of the vaccine warrant increased use of antivirals by hospitalists.

Dr Chang: Would you treat a 12-month-old previously healthy child admitted to the hospital with bronchiolitis, dehydration, and mild increased work of breathing if the testing for influenza was positive?

Dr Quinonez: Obviously if we followed the CDC recommendations, the answer would be yes. But let's think about this, ignoring this recommendation for the sake of argument. Would I start a medication on a child who is already ill that may have some effect on duration of symptoms but also may have significant risk of causing vomiting in a child who is already dehydrated? The risk outweighs the benefit in my mind, particularly if the child has been sick for over 48 hours, in which case the evidence for benefit is even weaker for antivirals. Most kids who present to the hospital with bronchiolitis are usually between day 3 and 5 of illness. So my evidence-based risk assessment, harm vs benefit answer, would be no.

Dr Garber: I agree with Ricardo. We know that bronchiolitis is a self-limited disease with basically zero mortality in this patient population. This particular child was admitted for dehydration, a minor complication of this disease that can be treated with simple interventions (increased nasal suctioning, small frequent feeds, or, if necessary, intravenous or nasogastric fluids). The average length of stay for such a child is quite short.

As an experienced hospitalist, I would say that a previously healthy child with bronchiolitis older than 1-2 months of age admitted solely for dehydration with no oxygen requirement usually goes home the following day, often in less than 24 hours. If the child has been ill more than 48 hours (and a close reading of the evidence suggests that most benefit is seen in patients treated within 24 hours), I see no evidence that oseltamivir will benefit this child, though it might induce emesis, which has a high likelihood of prolonging the hospitalization, increasing the potential for an adverse event from the hospitalization itself.

If the child has been ill fewer than 24 hours, I think a discussion with the family is appropriate. Prescribing oseltamivir may decrease his duration of symptoms by about a day and may prevent an otitis media. Side effects and costs should also be discussed with the family, and a shared decision to prescribe or not prescribe should be made, with anticipatory guidance about stopping the medication if side effects occur.