Listen to my interview about anti-flu drugs with broadcaster Andrew Easton on BBC local radio.
Pandemics of flu in past decades have caused serious complications and high death rates, with the 1918 'Spanish flu' the most serious example. This has lead to major efforts by the pharmaceutical industry to develop effective antiviral treatment for flu, to complement prophylaxis from immunisation.
Efforts to find drug treatments are important as flu mutates regularly so that immunisation may not protect from new strains of the virus.
During the 2009 swine flu pandemic, large amounts of antiviral drugs were prescribed and stockpiled.
A Cochrane review has now reported marked variability in the quality of data of clinical trials of drugs used to prevent and treat influenza and disappointing outcomes from higher quality studies. Many studies were reported to include important potential bias. Reasons included obvious differences in colour between placebo and active treatments in randomised controlled trials.
Serious outcome measures were often based on self-reporting, for example not using X-rays to confirm a presumptive diagnosis of pneumonia. And unpleasant side-effects of the treatment occurred in from ~1 in 25 to 1 in 90 patients, in particular headaches, vomiting and psychiatric abnormalities.
For the serious complication pneumonia, for studies in which X-ray confirmation was used, there appeared to be no benefit from active anti-viral treatment on pneumonia severity.
Although there was a reported ~10% reduction in severity of symptoms in adults and 1 in 6 shorter symptoms in children when drugs were given after the diagnosis, higher risk children - those this asthma, did not appear to experience this benefit on duration of symptoms.
There is now a presumption that alternative simpler and cheaper remedies such as paracetamol (acetaminophen) might be just as effective for symptom relief. The international incidence of flu is so high that trials to confirm that important question would not be difficult to complete rapidly.
Any such trials should use robust outcome measures, including recording admissions to hospital, confirming that flu diagnosis is correct (ie illness not due an infection such as RSV which is resistant to current flu antiviral agents), and serious complications are confirmed using tests such as X-rays.
A serious underlying issue arising from work by these Cochrane authors, and other assessments of safety and effective of drugs, was a delay of around 5 years in provision of trial data by one of the companies involved. Prompt access to all data is vital to ensure that the clinical and cost-effectiveness of drugs can be validated.
There are now important efforts underway from NGOs such as AllTrials, regulatory agencies and professional societies to improve transparency of access to company data by independent researchers.