A collective shudder of dismay rattled wine glasses on middle-class dining tables this week when a report labelled some of Britain’s most affluent towns as sozzled dens of “hazardous drinking” iniquity.

Middle-class, middle-aged drinkers, according to a report commissioned by the Department of Health, had been consuming so much wine for so long that many were putting their health “at significant risk”.

Cue the sound of corks returning to half-drunk bottles across the land. Had the users known how safe drinking limits came into being, they might just have poured another glass.

When the report defined any man who drinks more than 21 units of alcohol a week, or woman who consumes more than 14, as a hazardous drinker, the authors were relying on limits that have been set in stone for the past 20 years.

Yet these guidelines have no basis in science. Rather, in the words of a member of the committee that drew them up, they were simply “plucked out of the air”.

The safe limits were introduced in 1987 after the Royal College of Physicians produced its first health report on alcohol misuse. In A Great and Growing Evil: The Medical Consequences of Alcohol Abuse, the college warned that a host of medical problems – including liver disease, strokes, heart disease, brain disease and infertility – were associated with excessive drinking. The report was the most significant study into alcohol-related disorders to date.

But Richard Smith, the former editor of the British Medical Journaland a member of the college’s working party on alcohol, told The Times yesterday that the figures were not based on any clear evidence. He remembers “rather vividly” what happened when the discussion came round to whether the group should recommend safe limits for men and women.

“David Barker was the epidemiologist on the committee and his line was that ‘We don’t really have any decent data whatsoever. It’s impossible to say what’s safe and what isn’t’.

“And other people said, ‘Well, that’s not much use. If somebody comes to see you and says ‘What can I safely drink?’, you can’t say ‘Well, we’ve no evidence. Come back in 20 years and we’ll let you know’. So the feeling was that we ought to come up with something. So those limits were really plucked out of the air. They weren’t really based on any firm evidence at all. It was a sort of intelligent guess by a committee.”

On that basis, a nation’s drinking destiny was determined. The Government accepted the recommendation and 20 years later Professor Mark Bellis, director of the North West Public Health Observatory, which produced this week’s study, felt able to say that anyone exceeding the limits was “drinking enough to put their health at significant risk”.

That a host of epidemiological studies had filled the intervening years with evidence to the contrary seemed not to matter one jot. Most significant, perhaps, was a study carried out by the World Health Organisation in 2000.

The WHO’s International Guide for Monitoring Alcohol Consumption and Related Harm set out drinking ranges that qualified people as being at low, medium or high-risk of chronic alcohol-related harm. For men, less than 35 weekly units was low-risk, 36-52.5 was medium-risk and above 53 was high-risk. Women were low-risk below 17.5 units, medium between 18 and 35 and high above 36.

Seven years earlier, in 1993, a study of 12,000 middle-aged, male doctors led by Sir Richard Doll and a team at the Radcliffe Infirmary, Oxford, found that the lowest mortality rates – lower even than teetotallers – were among those drinking between 20 and 30 units of alcohol each week.The level of drinking that produced the same risk of death as that faced by a teetotaller was 63 units a week, or roughly a bottle of wine a day.

By 1994, five studies had been published which showed that moderate amounts of alcohol gave some degree of protection against heart disease. A year later, scientists at the Institute for Preventive Medicine in Copenhagen, who studied 13,000 men and women over 12 years, found that drinking more than half a bottle of wine a day – 50 units a week – cut the risk of premature death by half.

So what is the truth? Clarity is not aided by the fact that different countries use different quantities of alcohol to define a unit.

In Britain one unit of alcohol is 8 grams of pure ethanol. In Australia and Spain it is 10 grams, in Italy 12, in America 14, and in Japan 19.75. Translate the respective countries’ levels into British units and you find that, for men, Britain’s supposed safe weekly limit of 21 is more than Poland (12.5), but less than Canada (23.75), America (24.5), South Africa and Denmark (31.5) and Australia (35).

Some countries say that women should drink less than men, but others, including Canada, the Netherlands and Spain, make no distinction.

Christopher Record, a liver-disease specialist at Newcastle University, suggested that “it doesn’t really matter what the limits are”. “What we do know is, the more you drink, the greater the risk. The trouble is that we all have different genes.Some people can drink considerably more than [the limits] and they won’t get into any trouble.”

Explore Food & Drink

• Gordon Ramsay

• Heston Blumenthal

• Recipes

• Wine

• Real Food

• Drinks

• Anjum Anand

• Eating Out

• Video Recipes

Times Recommends

• Top Ten most searched for recipes
• Chalet Chef 2010: the winning menus
• Gin and pomegranate jellies