Oseltamivir, sold for years as Tamiflu and now widely available as a generic, was approved because it appeared to shorten influenza symptoms by about a day in otherwise healthy outpatients. That modest benefit, established in younger and less severely ill people, became the basis for enormous stockpiles and routine prescribing to the elderly and the hospitalized, the very patients for whom the evidence was thinnest. A randomised trial of critically ill flu patients has now reported that the drug not only failed to help them but was associated with substantially higher mortality.
The REMAP-CAP investigators (link below), assigned 442 hospitalised, critically ill adults with influenza to oseltamivir or placebo on top of usual care. Patients were typically around sixty and already carried other illnesses. Over the next ninety days, roughly 20 percent of those who received the drug died, compared with about 14 percent of those who did not. After adjustment for baseline differences, the untreated group was older and sicker, the researchers concluded that oseltamivir more than doubled the risk of death. They estimated a 98 percent probability that the drug was actively harmful and stopped giving it in the trial. Deaths among treated patients more often involved progressive multi-organ failure; rates of Aspergillus infection were also higher. The authors suggested a plausible mechanism: oseltamivir inhibits neuraminidase, an enzyme present not only on influenza particles but on human cells, and may therefore blunt host immune responses when they are most needed.
The finding is a preprint, not yet a peer-reviewed journal article, and the sample is modest. Even so, it is the first gold-standard randomised comparison of the drug in the population that actually dies of influenza. For more than two decades the medicine was given to millions of such patients on the assumption that a small benefit in mild disease would translate into lives saved in severe disease. That assumption was never properly tested until now.
The historical record already contained warnings. Cochrane reviewers who obtained the full clinical study reports found that oseltamivir modestly reduced time to symptom relief but produced nausea and vomiting and offered little clear evidence that it prevented serious complications or transmission. Psychiatric and neuropsychiatric events, delirium, abnormal behaviour, hallucinations, were reported, particularly in children, and led Japanese regulators to issue cautions. Later observational work has been mixed; some large paediatric cohorts have associated treatment with fewer neuropsychiatric events, attributing those events to influenza itself rather than the drug. The new trial does not settle the paediatric question. It does raise a more urgent one about the elderly and the critically ill, the groups most likely to be prescribed the drug in hospital.
Roche sold roughly twenty billion dollars of Tamiflu before the patent expired. Governments bought mountains of it during the 2009 swine-flu scare. In the United States it is still dispensed to nearly two million people a year. If the REMAP-CAP signal is real, a medicine marketed as a shield against flu death has instead increased death among those already closest to it. Extrapolations to tens or hundreds of thousands of excess deaths over twenty-seven years remain speculative; they depend on how widely the drug was used in comparable patients and on whether the trial result generalizes. What is no longer speculative is that the pivotal evidence for the highest-risk group arrived a generation late.
A medicine that shortens a mild illness by a day can still be useful for some outpatients who start it early. That is not the same as a medicine that belongs in the intensive-care unit. The new data suggest oseltamivir may impair recovery precisely when recovery is most fragile. Until larger, peer-reviewed confirmation or refutation appears, the prudent course is to treat the drug as unproven, and potentially harmful, in critically ill influenza patients, rather than as the default antiviral it has long been assumed to be. The trial that should have been run at the beginning has finally been run. Its result is not the one the stockpiles were built to expect.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7172531
https://alexberenson.substack.com/p/urgent-tamiflu-kills