However, what is new since the last influenza season is incremental progress in opening the data on oseltamivir. As has been detailed extensively on the BMJ Tamiflu campaign, substantial doubts remain regarding the effectiveness and utility of broad use of oseltamivir for seasonal influenza.³ Independent investigators from The Cochrane Collaboration have persistently campaigned for Roche, manufacturer of oseltamivir, to provide access to full trial reports, including several early trials for which no data have been released. The overwhelming concern from this group of investigators is the data currently available for review incompletely describe the adverse effects of neuraminidase inhibitors and the quality of reporting of complications is low. Extensive meetings between the Cochrane group, Roche, and another Roche-funded reanalysis group have resulted in additional data available for analysis.

The first such updated publication provides a mixed picture regarding the utility of oseltamivir. The authors report that oseltamivir use in symptomatic adults reduced average symptom duration from 7 to 6.3 days and that use of oseltamivir as influenza prophylaxis reduced risk of influenza transmission at a rate of one case prevented for every 33 treated. However, the use of oseltamivir did not exhibit a significant effect reducing subsequent influenza complications or hospitalizations while resulting in an excess of nausea, diarrhea, cardiac adverse events, and neuropsychiatric effects. The Cochrane group concludes the use of oseltamivir for treatment of influenza must consider potential benefits and harms of treatment on an individual patient basis.⁴

The CDC, however, prefers to justify its recommendations mostly on observational data reflecting decreased mortality following treatment with neuraminidase inhibitors.⁵ CDC representatives acknowledge the trials may demonstrate little benefit to use in healthy individuals, but they note high-risk patients are underrepresented and the observational reports from pragmatic use may be more informative for such populations. Unfortunately, many of these observational reports are from investigators funded by Roche, have hotly debated statistical methodology, and are subject to the many biases inherent to such retrospective investigations.⁶

What’s the bottom line? We still don’t have good evidence describing the efficacy of oseltamivir. An otherwise healthy patient with influenza-like symptoms may receive a trivial half-day reduction in symptoms associated with a weeklong illness at the expense of the cost of medication and risk of adverse medication effects. If this season again is primarily H1N1 or a new highly pathogenic strain, it is probably reasonable to offer the drug to high-risk individuals while communicating that the benefits are suspected but unproven. Finally, for hospitalized patients with severe illness, the potential benefits almost certainly outweigh the small incremental costs of oseltamivir relative to those of hospitalization, and observational evidence, flawed as it may be, supports its use.