A new approach to preclinical tests could help avoid a repeat of the disastrous trial of an antibody in a North London hospital in 2006 which left six healthy volunteers hospitalised.

This is the view of Stephen Poole, who was speaking at the British Pharmacological Society’s winter meeting in Brighton, described the incident at Northwick Park as "the most obvious setback for medicines testing since thalidomide". Shortly after being dosed with the compound, TGN1412, developed by German biotechnology company TeGenero as a treatment for chronic inflammatory diseases and leukaemia, the men experienced multi-organ failure.

Now Dr Poole, and his colleagues from the government-funded National Institute for Biological Standards and Control (NIBSC), say they have successfully reproduced the devastating reaction suffered by the volunteers using human cells in vitro. The Northwick Park disaster resulted in the UK Medicine and Healthcare products Regulatory Agency suspending all clinical trials of immunotherapy drugs and commissioning the NIBSC to investigate why both in vitro human cell tests and in vivo animal tests failed to predict the human immune system's response to the drug.

"With hindsight, testing this drug in man was a mistake, but at the time the standard required preclinical tests failed to predict the effects it would have on the six volunteers," said Dr Poole. His team have identified new preclinical in vitro testing methods for immunotherapies that should help prevent any repetition of the disastrous events witnessed at Northwick Park.

These measures include ensuring that such drugs are not tested solely on immune (white blood) cells in isolation. The NIBSC group has shown that having a mixed human cell culture of white blood cells and endothelial cells – the cells that line blood vessels – is a much better indicator of how this type of drug will react in vivo.

The NIBSC group has also developed a technique that dries the drug onto a plastic surface, rather than testing it on cells as a solution in water, which has proved to be a far more reliable indicator of how the drug will react in the human body. By Nick Mason