It's possible that the Tamiflu pandemonium may be peaking for most emergency and primary care physicians because the dubious benefits of oseltamivir have now been well documented, discussed, and disseminated. (See box.) But there's also a new kid on the block (baloxavir marboxil, AKA Xofluza) that is all too likely to be jammed down our collective throats, displacing its controversial cousin.
That said, it seems we never go too long before another article comes along touting the life-saving properties of this wonder drug. It's not that the medicine doesn't work, its proponents claim, it's that we're giving it to the wrong population or with the wrong timing or in the wrong amount. We end up with endless speculation on how the mechanism of action for this medication might be more or less effective against various subtypes of influenza, despite the fact that the authors of the definite Cochrane Review about this subject have written that “the influenza virus-specific mechanism of action proposed by the producers does not fit the clinical evidence.” (Cochrane Database Syst Rev 2014 Apr 10;:CD008965; http://bit.ly/2Hdwgx5.)
The newest entry in the saga to find the elusive One True Cohort™ for Tamiflu is an article by Lytras, et al., a retrospective cohort of 1330 very sick patients admitted to Grecian ICUs over eight flu seasons starting in 2010. (Clin Infect Dis 2019 Feb 7. doi: 10.1093/cid/ciz101.) By very sick, I mean very sick: All the patients were intubated, had laboratory-confirmed influenza, and were treated with oseltamivir. Nearly half (46.8%) of the patients in this study died in the ICU. This publication has generated some pro-oseltamivir headlines, so it's probably worth taking a look at the results for patients treated early (48 hours or less from symptom onset ) or late (48 hours or more after symptoms):
- Death in the ICU (primary outcome): No difference
- Death in the ICU for A type H1N1 strain: No difference
- Death in the ICU for A type H3N2 strain: No difference
- Death in the ICU for B type influenza: No difference (obviously)
- Increased cause-specific hazard for discharge for A type H3N2 strain: csHR 1.89, 95% CrI 1.33-2.70
The first (and second and third) time I looked at those results, I thought, “Well, that looks like another swing and a miss for Tamiflu.” The authors, however, arrived at a slightly different conclusion: “[I]n the absence of randomized trials, our study provides strong and important new evidence about oseltamivir use in critically ill influenza patients.”
The Bottom Line
I turned to a colleague to seek some clarity. Brett Faine, PharmD, an emergency medicine clinical pharmacy specialist and an assistant professor of pharmacy practice and science at the University of Iowa College of Pharmacy, has a passion for evidence-based clinical practice and, in a related development, a long-standing skepticism of oseltamivir and the way it has been marketed and recommended around the world. My first question to him was whether I was alone in my confusion about the conclusions from this paper. “No, I think it's important to note that this was a negative study,” he said. “There was no difference in the primary outcome. Period. Overall, early oseltamivir didn't appear to save any lives, even in this very sick cohort.”
How did we go from there to a recommendation to keep giving Tamiflu to critically ill patients? “They are basing this entirely on a finding from a subgroup analysis [only 20.8% of patients], which found that fewer patients with the A/H3N2 subtype died in the early oseltamivir group [33.7% vs 48.4% respectively, p=0.029],” Dr. Faine said.
Were patients at less risk of dying in the ICU if they got early Tamiflu and had H3N2? “No,” he said. “Actually there was no difference in deaths in the ICU for these patients. The benefit was entirely driven by an increase in the number of patients in this group who were discharged from the ICU.”
I'm still confused, so I asked him to explain it once more. “The cumulative hazard analysis revealed that among the A/H3N2 patients, the hazard of death was no different, but the hazard of ICU discharge, a good thing in this case, was significantly increased,” he said. “The authors argue that the instantaneous risk of dying is not [affected] by Tamiflu but that the hazard of discharge from ICU translated into lower mortality. They discuss that this is likely due to the mechanism of action—inhibition of viral replication and neuraminidase, but I would just note that the Cochrane authors call into question whether oseltamivir even works this way at all.”
Are there other reasons to be skeptical of these findings? “Where to start?” Dr. Faine said. “It was a negative study with a positive subgroup analysis. The study is limited based on the usual stuff that plagues observational studies, but also the fact that we do not have information on dosage or duration, so we have no idea how much of the medicine was given, for how long, or if it was discontinued. We also get no data on adverse events. It took a lot of statistical massaging to come to the conclusion that Tamiflu was effective in A/H3N2. And we have no data on overall mortality in these patients. We have no idea what happened to them once they left the ICU, no data on hospital discharge, inpatient mortality, or survival at any time post-discharge.”
Bottom line? “In my opinion,” Dr. Faine said, “the conclusions made in this article, specifically that we give Tamiflu to all sick influenza patients, is not supported by the study findings.”
To his credit, Theodore Lytras, MD, PhD, also acknowledged some room for doubt in an interview on the topic. “We would not describe the results of our study—or any single study, no matter how well-designed and analyzed—as a basis for ‘firm conclusions,’” he said. “[W]e believe our results will need to be replicated in further studies, ideally randomized.” (MD Feb. 22, 2019; http://bit.ly/2EntfXR.)
We should remember that almost all of the literature purporting that oseltamivir decreases mortality has been related to H1N1, and if there's one thing that this study makes abundantly clear, it's that Tamiflu appears pretty worthless for the H1N1 cohort.
Read Dr. Runde's first article, “Still Prescribing Oseltamivir?” at http://bit.ly/JustSayNoToTamiflu.
Dr. Rundeis the assistant residency director and an assistant professor of emergency medicine at the University of Iowa Hospitals and Clinics, where he serves as co-director for the associate fellowship in medical education. He creates content for and is a member of the editorial board forwww.TheNNT.com, and is a content contributor forwww.MDCalc.com. Follow him on Twitter @Runde_MC, and read his past articles athttp://bit.ly/EMN-ReasonableDoubt.