The effects of antiviral drug Tamiflu - which was stockpiled by the Government at the height of the swine flu pandemic - are still uncertain, researchers said as they accused pharmaceutical giant Roche of withholding data.

A 2009 study in the British Medical Journal found the drug may cut the length of time people have symptoms but found no clear evidence that it prevented complications such as pneumonia.

Experts from the Cochrane Collaboration conducted a review of 20 existing studies.

They found Tamiflu "did not reduce influenza-related lower respiratory tract complications" but did induce nausea.

Evidence of other adverse reactions to the drug were "possibly under-reported", they said.

The study criticised the evidence available from Roche and said the firm had not been able "unconditionally" to provide the information needed for the study.

At the height of the swine flu pandemic, the UK Government had contracts in place for 50 million doses of Tamiflu - enough for 80% of the population. Tamiflu is still available for use when flu is circulating at high levels.

A new report from the Cochrane Collaboration said the question of how well Tamiflu works remains unresolved owing to Roche's continued refusal to provide full access to its data.

Following the 2009 study, Roche did provide 3,195 pages of study reports for 10 treatment trials of Tamiflu but these were incomplete, the experts said.

A statement from Roche said: "Roche has made full clinical study data available to health authorities around the world for their review as part of the licensing process. It is the role of global health authorities to review detailed information on medicines when assessing benefit/risk."