Influenza and oseltamivir

Journal Club Synopsis, January 2016.  Oseltamivir-Prescriptions and Profits.

Many thanks to Branka Milicev for tasty Italian food and beautiful views from her south loop penthouse!

The Tamiflu controversy heated up in 2014.  Turns out Roche, manufacturer of oseltamivir, had suppressed a number of unpublished trials.  A group of researchers went after Roche for several years, and after much public shaming, Roche finalized released these data.  Studies incorporating the unpublished data were published in the April 9, 2014 issue of BMJ as well as our included Family Practice article.  The authors of the BMJ articles also updated the pertinent Cochrane review in 2015, which covers the use of all neuraminidase inhibitors in preventing and treating influenza.

Article 1: (Regan, Navarette)

Ebell MH, et al. Effectiveness of oseltamivir in adults: a meta-analysis of published and unpublished clinical trials.  Family Practice 2013; 30:125–133 doi:10.1093/fampra/cms059

This meta-analysis includes published and unpublished double-blinded, placebo-controlled RCTs of oseltamivir in adults with suspected or confirmed influenza that reported the primary outcomes of duration of symptoms, complications, or hospitalizations.  All of these studies were limited to patients presenting within 36 hours of symptom onset.  Pregnant patients were excluded by most of the reporting articles.  Authors were granted access to trial registries maintained by the drug manufacturer.  Eleven studies (3 published, 8 unpublished) with a total of 4769 patients met inclusion criteria. 

For the intention to treat (ITT) population, the mean reduction in duration of symptoms was 21 hours.  There was no significant reduction of symptom duration in the elderly and in patients with chronic disease, two populations felt to be at higher risk.  One study stratified by time of symptoms:  reduction of symptoms by 29 hours if oseltamivir was given within 24 hours of symptom onset, 15 hour reduction if given between 24-36 hours of symptom onset.

Overall, there was no difference in the likelihood of hospitalization and no reduction in the rate of infectious complications requiring an antibiotic when acute bronchitis was excluded.  Authors excluded acute bronchitis as current guidelines do not support the use of antibiotics for acute bronchitis.  The risk of pneumonia was reduced in the population with PCR confirmed influenza infection (NNT = 111), but not in the general ITT population.  There was only one death reported, in the placebo group. 

Downsides to Tamiflu?  Cost is about 100$.  Side effects of oseltamivir were not specified in this meta-analysis, but other reports state rates of nausea and vomiting between 5-10%.  There are less common neuropsychiatric and renal complications as well as headache.  An additional concern is the expected increase in drug resistance with liberal oseltamivir use.  Authors also comment on the questionable motivations of the manufacturer not publishing multiple studies, and the break in the public trust when patients are recruited and enrolled in studies whose results are never used for the intended purpose of advancing medical care.

Limitations of the meta-analysis include the limited reporting of complications in the ITT population.  Existing literature includes only small numbers of hospitalizations and serious complications, which limits comparisons, especially for high-risk patients.

From our presenters: they are less likely to prescribe Tamiflu.   One more day of flu has different meanings to different people, depending on their income, work situation, future exposures, comorbidities.   

Bottom line:  In this large meta-analysis of published and unpublished studies, there is no evidence that oseltamivir reduces the likelihood of hospitalization or acute infectious complications.  There is a small reduction in the risk of pneumoniabut this was only seen in lab confirmed influenza (NNT =111).  There is a reduction of about one day in duration of symptoms if oseltamivir is given within the first 36 hours of illness, but treatment is associated with side effects of nausea and vomiting.


Article 2: (Einstein, Ede)

Muthuri SG et al. Eff ectiveness of neuraminidase inhibitors in reducing mortality in patients admitted to hospital with influenza A H1N1pdm09 virus infection: a meta-analysis of individual participant data. Lancet Respir Med 2014;2: 395–404.

So, if oseltamivir doesn’t help prevent infectious complications or hospitalizations in outpatients, does it at least help in sicker patients with flu who require hospitalization? 

This is a meta-analysis of retrospective, observational data from the 2009-2010 Influenza A H1N1 pandemic.  It included data for 29,234 patients from 78 studies worldwide, who were admitted to hospital with clinical or lab confirmed influenza A H1N1.  Primary outcome was the association between neuraminidase inhibitor treatment and mortality, adjusted for treatment propensity and confounders.  Although the authors state in their methods that they attempted to include both observational studies and RCTs, in their conclusions they refer to using retrospective, observational data and in the appendix, there is no further clarity regarding the methodology of studies included. 

In adults, neuraminidase inhibitor treatment was associated with mortality risk reduction, adjusted odds ratio 0.81; 95% CI 0.70-0.93.  Treatment within 2 days of symptom onset vs. late treatment, treatment within 2 days vs. no treatment, and treatment of pregnant women and critically ill adult patients were all associated with even greater mortality risk reductions.  If treatment was started within 2 days, NNT (to save a life of an admitted pt) was 31.

Combining all subgroups, there was no mortality benefit with treatment delayed more than 2 days after symptom onset; however there was still benefit of delayed treatment in patients admitted to ICU.  The association with reduced mortality was not statistically significant in pediatric patients. 

What are the problems with this study?

First, funding was by Hoffman –La Roche, makers of oseltamivir, and the main author was a paid consultant for Pharma and former employee of Roche.  As seen from the first article, there have been big issues with suppressed data and overstatement of treatment effect with this drug.

The data were retrospective and observational, implying potential for lots of residual confounders and missing information.  Although 401 centers were contacted to contribute, only 78 centers were included.  Finally, the quality of the included studies has been described as “medium to poor,” with “clear evidence of unadjusted biases.” (

Bottom line:  In this meta-analysis with significant limitations, oseltamivir treatment within 2 days of symptoms was associated with a mortality reduction in hospitalized adults, with the greatest benefit seen in pregnant women and critically ill patients.


Article 3: (Hart, West)

Lenzer J. Why aren’t the US Centers for Disease Control and Food and Drug Administration speaking with one voice on flu? BMJ 2015;350:h658 doi: 10.1136/bmj.h658

Ok, I’ll get a little political.

From “Tamiflu, a Nice Little Earner:-BMJ, 2014.

Since its launch in 1999, oseltamivir has generated cumulative sales in excess of $18 billion for Roche. Half of the total expenditure was by governments and companies around the world for stockpiles for pandemic preparations. The US alone spent more than $1.3 billion buying a strategic reserve of antivirals.  Most have never been used, and today the US stockpile consists of more than 65 million treatments. 

Back to the Lenzer’s editorial:

The FDA disagrees with the CDC  “Take 3” campaign that claims oseltamivir “saves lives.”  The FDA has stated that oseltamivir “has not been proven to have a positive impact on the potential consequences (such as hospitalizations, mortality, or economic impact) of seasonal, avian, or pandemic influenza.”

CDC director Thomas Friden has made expansive claims about the benefits of neuraminidase inhibitors for non-admitted patients, including the implication of a general mortality benefit.  This approach has been called irresponsible, and an appeal of emotion instead of science.  Oseltamivir costs around 100$, which for many patients is a substantial cost for a marginal benefit.  Sending large numbers of otherwise healthy people to the ED who have influenza but who won’t actually benefit from oseltamivir may actually increase the spread of the virus around crowded ED waiting rooms.

The “CDC Foundation” provides funding to the CDC.  The CDC received a directed donation of $198,000 from Roche (maker of oseltamivir) via the Foundation for the “Take 3” flu campaign.  Step 3 of “Take 3” encourages people to “take flu antiviral drugs if your doctor prescribes them.”  Overall, the CDC receives millions of dollars annually from Big Pharma, channeled through the CDC Foundation.  The reputation of the CDC as being free of industry bias is viewed by some as now compromised.

The CDC cites the Lancet’s 2015 Dobson meta-analysis as demonstrating a reduction in hospitalization after outpatient treatment with oseltamivir.  This study was indirectly sponsored by Roche, and according to Jeanne Lenzer, “all four researchers received industry funding either directly or through industry donations to organizations that directly funded the study (so-called “pass-through” money).”  The Dobson paper’s conclusions are far different than the conclusions from Journal Club’s meta-analysis by Ebell and the 2014 BMJ systematic review by Jefferson, both of which included previously unpublished data from Roche.

Even after the 2014 articles in the BMJ and Family Physician, the CDC’s recommendations regarding neuraminidase treatment for influenza are unchanged.

Really though, if you want the dramatic interpretation, please tune in to youtube and our very own thespians Elise Heeringa and Nathan West.  This was forwarded to Jeanne Lenzer at the BMJ (author of the editorial), who was very impressed J

y Journal Club bottom line:  for outpatients, there’s just no data to support treatment with neuraminidase treatment to prevent serious complications, hospitalization, or death.  At best, for outpatients, if treated within 36 hours of symptom onset, you’ll get 21 hrs of decreased illness, for a 100$ drug with up to 10% incidence of N/V.  For adult inpatients, if oseltamivir is started within 2 days, there is likely a mortality benefit.  For admitted kids, overall data says mortality same with/without oseltamivir.  However, there is a December 2013 Pediatrics study (Louie et al) of kids admitted to the ICU that demonstrated a mortality benefit.