THE recent admission by pharmaceutical giant GlaxoSmithKline's vice-president of genetics that 90 per cent of their drugs only work in 30-50 per cent of people is a tragic consequence of the continued use of animal tests for assessing the safety and effectiveness of drugs.

Not only do these unreliable experiments lead to drugs which don't work, they also enable dangerous products to flow on to the market.

Among the many drugs that were supposedly proved 'safe' in animal tests are treatments for ulcers, asthma and arthritis -- that have gone on to demonstrate serious, sometimes fatal, adverse reactions in people.

A survey of 116 drugs withdrawn between 1960 and 1990 because of their side effects showed that in only 11 cases did animal tests even begin to hint at the problems that ultimately occurred in humans.

Misleading data from animal tests also delays real cures reaching people. After its discovery, penicillin was initially disregarded because it was shown to kill rabbits.

Only when given to a dying human patient as a last resort many years later were its therapeutic qualities recognised.

Computer programmes, cell cultures and micro-dosing (giving patients a tiny amount of a drug to be tested) all provide accurate information which is relevant to humans.

Only when these methods become the industry norm, rather than experimenting on mice, rats, cats, dogs and monkeys, will medical research really progress.

CLAUDIA TARRY, Campaigns Officer, Animal Aid, Tonbridge, Kent.