A little of what you fancy may be extremely good for you. But although the news may be extremely palatable for the regulars in the nation’s saloon bars it is not so welcome for those who have to educate the public about health risks.

The latest research shows that a few glasses of beer or wine every week could help to reduce the risk of dying from heart disease – one of the commonest causes of death in the Western world. But alcohol can cause a string of other diseases including cirrhosis of the liver, cancers of mouth and throat, hypertension and strokes. For public health educators, concerned to convey a simple message, the picture is now about as clear as a murky glass of beer.

The now-familiar message from the Health Education Authority is that the sensible limit is 21 units a week for men, and 14 for women. However, this new research is forcing a rethink. A working party from the Royal Colleges of Physicians, General Practitioners and Psychiatrists aims to publish their view by the end of the month. The government is also to issue a statement of ‘clarification’ this autumn.

They are all faced with a dilemma: any acknowledgment of alcohol’s good points could encourage everyone to drink more. This might increase the number of excessive drinkers, and the health and social problems, including crime and violence, associated with heavy drinking.

The research does not point to an ideal level of drinking. But Richard Doll, emeritus professor of medicine at Oxford University says that medical thinking has changed over the past five years and it is now generally accepted by everybody who has looked at the data that alcohol has some beneficial effect. He says: ‘What is difficult to assess is the extent to which this counteracts its harmful effects. Current evidence suggests that it probably does. The question is: up to what level of drinking?’

One of the most startling new findings suggests that the answer to this question may be as high as 42 units a week. A Danish study of 13 000 men and women, published this year, found that teetotallers were more likely than drinkers to die of any cause, including those causes that are not related to excessive alcohol consumption. People who drank between 1 and 6 units a week were least likely to die, but the risk only rose significantly once people were drinking a hefty 42 units a week or more.

When the risk of death is plotted on a graph against alcohol consumption, these results give a U-shaped curve. In itself, this is not new. The shape of this curve was first discovered in the 1920s. But for many years doctors believed that the death rate of teetotallers was relatively high because they included substantial numbers of reformed drinkers, whose bodies had been already irreparably damaged by alcohol.

Unhealthy teetotallers

This argument has been steadily unravelling over the past few years, helped along by studies such as that by Rodney Jackson at the University of Auckland, which demonstrated that teetotallers who were reformed drinkers have a lower risk of coronary heart disease than lifelong abstainers.

The Danish study also looked for signs of alcoholic liver disease among abstainers, which would be expected if some had stopped because drinking was making them ill. No significant signs were found, providing further evidence that people who drink really are healthier than those who do not.

American research published late last year suggests a possible reason for the beneficial effect of alcohol. Doctors based at Havard Medical School in Boston, Massachusetts examined 340 people who had just had a heart attack and compared them with healthy people of a similar age. They found drinkers had a lower risk of suffering a heart attack than nondrinkers. Biochemical analysis of blood samples demonstrated that drinkers had higher levels of high-density lipoproteins which are known to protect against heart disease.

Researchers led by Serge Renaud, nutritionist and epidemiologist at a unit in Lyon that is part of Inserm, the French medical research council, believe that alcohol may also reduce the likelihood of a blood clot forming by acting on cells called platelets. A clot begins to form when platelets clump together. People who drink appear to have platelets that are less ‘sticky’ than those who do not. They have shown that farmers in Var in the south of the France drink more than twice as much alcohol as those in Stranraer, Scotland, but their platelets are less likely to begin to clot.

However, Morten Gronbaek, research fellow at Copenhagen’s Institute of Preventive Medicine, who led the Danish study stresses his research does not show that teetotallers would benefit from taking up drinking. He says that his work should not be used as evidence to change health guidelines. Although the research showed that the increase in mortality was not statistically significant until people drank more than 42 units a week, deaths started to rise at lower levels (see diagram). ‘It’s an estimation of risk. It wouldn’t tell you to change the recommended limit for drinking,’ he says.

This highlights the difficulties faced by public health educators. Since there is no threshold beyond which alcohol suddenly becomes harmful, recommended levels of drinking are arbitrary. Until the Health Education Authority laid down its limits in 1987 the Royal College of Psychiatry used to advocate a limit of 56 units a week for men.

Griffith Edwards, professor of addiction behaviour at the Institute of Psychiatry in London, says different diseases have different links with alcohol. Cirrhosis and alcohol has an exponential link, with the disease rising sharply among people who drink more than two to three drinks a day. Cancers tend to increase in direct proportion to alcohol consumption. Death from heart attacks, however, gives a J-shaped curve, or a reduction in mortality followed by an increase.’ We’re looking at a trade-off between different curves. We can’t talk bone-headedly about safe limits. It’s only relative risk.’

This ‘best guess’ approach does not make for a clear-cut public health message and medics are cautious about changing their advice. All the more so, because another strand of research suggests that the population as a whole might be healthier if the average consumption of alcohol was reduced.

Michael Marmot, professor of epidemiology at University College Medical School in London, one of the leaders of a major international study called Intersalt, which looked at almost 10 000 people around the world, found that those who have more than three drinks a day have significantly higher blood pressure than those who drink less. He says this shows there is only a ‘narrow window’ in which alcohol may protect against heart disease.

One of the key findings of this study is that the mean level of consumption in a population is closely correlated to the prevalence of heavy drinking. Any campaign which encourages people to drink runs the risk of increasing the mean level of consumption, says Marmot. If this happens it will probably increase the prevalence of heavy drinking, and with it, the serious health and social problems related to alcohol. Marmot says: ‘We are very reluctant to do anything which might increase heavy drinking.’

But John Duffy, a statistician at the University of Edinburgh’s Alcohol Research Group is dismissive. ‘It’s absolute madness. Of course there’s a correlation. It would have to be a strange world if a population with more heavy drinkers didn’t have a higher mean level of consumption.’ He does not believe that reducing average consumption is the way to tackle heavy drinking. He says: ‘If you want to reduce the population of heavy drinkers, you have to persuade heavy drinkers to drink less.’

Duffy says the British guidelines on sensible drinking are fair as a description of levels at which people are unlikely to run into difficulties. ‘What went wrong is that they got transformed in the public imagination to limits. And the problem with it is that it overemphasises the units of consumption and underemphasises the way drinking takes place.’

Binge drinking is known to be harmful. The Intersalt study demonstrated that the effect on blood pressure was greater in episodic drinkers – those who got plastered on a Saturday night – than those who drank a similar amount but paced their drinking throughout the week. Binge drinkers are also most likely to get drunk and be involved in accidents and violence.

Britain is unusual in suggesting weekly totals. Health authorities issue guidelines throughout northern Europe and America, but their advice focuses on drinks per day or drinks per occasion.

Juhani Lehto from the WHO’s alcohol, drugs and tobacco unit, says that in parts of Scandinavia, many people drink only at the weekend. ‘A weekly limit would be dangerous if people took it all in one go.’ However, France and Italy do not set specific limits. ‘It’s a different culture and philosophy. In southern Europe, the general feeling is that states should not give as much advice as we do in northern Europe,’ says Lehto.

Nevertheless, researchers in Britain believe that the recommended levels are useful. Paul Wallace, professor of primary health care at London’s Royal Free Hospital School of Medicine says the new research on heart disease does not fundamentally change our knowledge on safe drinking. He says the limits are intended as a definition of risk rather than a threshold. But he believes that the medical profession must acknowledge the new evidence in order to make a responsible interpretation of it. ‘We should respond to the debate and the points being made in the press,’ he says.

How much more information can reasonably be given is a thorny issue. One of the complicating factors is that different groups in society have different risks. For 50-year-old men drinking will be beneficial, because reducing the risk of heart disease will outweigh any increased risk of cancer. But 30-year-old women are better off not drinking because the reduction in their much lower risk of heart disease does not quite match the increase in the risk of breast or bowel cancer.

According to John Kemm, public health consultant at South Birmingham Health Authority this kind of information is too complicated to be used in a public campaign. But Martin Plant, professor of sociology alongside Duffy in Edinburgh, does not see any dangers. ‘It might make the message more complicated but I don’t think it’s honest to dismiss the evidence on the grounds that it might be confusing,’ he says. ‘People have to be allowed to be responsible for their own actions.’

Helen Saul is a freelance medical writer.

And yet, almost 40 years after details of the technique were first published,
doubt is being cast over the wisdom of ultrasonic screening. Last year,
four studies suggested that, at best, routine scanning makes no overall
difference to the health of babies. At worst, it could actively harm them.
The reports support scanning when there is a clinical reason for it, such
as if a woman is bleeding, if she smokes, or has had a slow-growing fetus
in a previous pregnancy. But they suggest that we should reconsider the
practice of scanning virtually every pregnant woman.

The first study to make waves came from the US. The RADIUS trial – for
routine antenatal diagnostic imaging with ultrasound – followed more than
15 000 women who had been sorted at random into two groups. In one, all
women were routinely scanned twice. In the other, a scan was carried out
selectively, only when doctors identified a medical reason. The results
show that the problems suffered in the two groups were almost identical.
There was no difference between them in the birth weight of babies or the
number born prematurely. Nor was there any difference in the number of babies
born with congenital malformations (The New England Journal of Medicine,
16 September 1993, p 821).

Australian work published in The Lancet a month later (9 October 1993,
p 887), also found that routine screening had no advantage over selective
scanning. But this study, which followed 2800 women, also linked frequent
scanning to low birth weight babies. The team, led by John Newnham from
King Edward Memorial Hospital in Perth, concluded that, while this finding
could be due to chance, it is plausible that frequent exposure to ultrasound
may have slowed fetal growth.

Researchers elsewhere found other possible harmful effects. A group
in Norway studied children of eight and nine years and found that those
who were exposed to ultrasound in utero were more likely to be left-handed.
Kjell Salvesen at the University Medical Centre in Trondheim, and his colleagues,
also admit their finding could be chance, but say it might also be ‘a sensitive
index of subtle changes in the development of the brain’ (British Medical
Journal, 17 July 1993, p 159).

In Canada, James Campbell and colleagues from the University of Calgary
found children with delayed speech were twice as likely as normal controls
to have been exposed to ultrasound in utero (Canadian Medical Association
Journal, 15 November 1993, p 1435). They concluded that, while potential
side effects are subtle: ‘Physicians might be wise to caution their patients
about the vulnerability of the fetus to noxious agents.’

Gaps in knowledge

None of these results are devastating in isolation, but together they
raise serious doubts about the procedure. Also published last year was a
meta-analysis of the only four trials carried out around the world in which
women were randomly allocated to a test or control groups under strict criteria.
Routine screening did not come out well in this analysis either. It neither
increased the number of live births nor reduced the death rate around the
time of birth. It did find that early abortion of malformed fetuses reduced
the number of fetal deaths later in pregnancy and the number of children
born with malformations. But this benefit flowed mainly from a single study,
called the Helsinki ultrasound trial. The high rate at which malformations
were detected in this trial combined with Finland’s liberal abortion law
to produce exceptional results.

Writing in medical journals, critics have attacked the American and
Australian research. They argue that the RADIUS study suffered from insensitive
scanning: it picked up only 17 per cent of gross malformations compared
with 50 per cent in the Helsinki trial. And the Australians subjected women
to five scans, which is more than most women receive.

But if the studies fail to provide all the answers, they expose big
gaps in our knowledge. Years after routine ultrasound scanning became a
recognised part of pregnancy care in much of the Western world, we cannot
answer basic questions such as is routine scanning beneficial? And could
ultrasound be harmful?

One of the strongest objections to routine scanning is the problem of
false positive results – when women carrying healthy fetuses are told there
may be a problem. The obvious advantage of allowing a woman carrying a severely
malformed fetus to have an early termination must be balanced against the
distress caused to another woman given false information.

Three women in South Wales spoke last month of their anguish and anger
after being told by medical staff that scans had shown their fetuses to
be dead. Later it emerged that the fetuses were all alive. Jim Nielson,
professor of obstetrics at the University of Liverpool, says there is not
only a psychological price to pay for such mistakes, but also a risk to
the pregnancy. When a problem is detected during a scan, women can be referred
to have an amniocentesis, which involves taking a sample of the fluid surrounding
the fetus. ‘There is a small but finite risk of miscarriage associated with
amniocentesis,’ says Nielson. ‘There is a risk of occasionally losing a
normal pregnancy.’

Heiner Bucher, registrar at the University Polyclinic in Basel and author
of the meta-analysis, says the only justification for using routine ultrasound
is to screen for malformations. But even in the best study, the Helsinki
trial, for every 2.7 malformations that were diagnosed correctly, 2.4 women
were given false positive results. ‘You prevent two in a thousand women
giving birth to a malformed fetus. At the same time, you disturb two in
a thousand women with this false positive information,’ says Bucher. ‘In
absolute terms, the results are not convincing about the benefit of routine
³Ü±ô³Ù°ù²¹²õ´Ç³Ü²Ô»å.’

Bob Fraser, senior lecturer in obstetrics at the University of Sheffield,
says ultrasound may just not be sensitive enough to be a screening tool.
‘If you screen the whole population, the number of cases which are positive
are few and the false positives are such that you end up investigating
and intervening in any number of pregnancies. The technique comes under
a cloud.’ But selective scanning can discriminate between real cases and
false positives quite well, he says. ‘You can’t necessarily extrapolate
from a test which is clinically useful in people at high risk, say 10 per
cent of the population, to 100 per cent,’ says Fraser.

It would also make sense to restrict scans to women who need them for
medical reasons, if ultrasound turns out to be harmful. But the evidence
of harmful effects is not conclusive. The designs of all three of last year’s
studies that found adverse effects linked to ultrasound could have been
better. The Australian study, for example, was designed to identify the
benefits of scanning. The link between low birth weight and multiple scans
was a coincidental finding, which does not carry the same statistical weight
as the results the trial was designed to generate.

Valerie Beral, an epidemiologist at Oxford’s Radcliffe Infirmary, says
any dramatically harmful side effects of ultrasound would have come to light
by now because so many children have been exposed to it in the womb. The
number of babies who are born blind or deaf, for example, is monitored and
any increase since ultrasound was introduced would have been picked up.
But more subtle or rarer side effects might not have been found.

Bernard Ewigman of the University of Missouri School of Medicine in
Columbia, and principal author of the RADIUS trial, wrote in 1989 that the
level of research into the safety of ultrasound is similar to that into
fetal X-rays before 1950. Only later did studies find that X-rays increased
a child’s risk of developing leukaemia by nearly threefold. Leukaemia affects
about 1 person in 5000. Ewigman wrote: ‘No current epidemiological study
of ultrasound exposure has an adequate sample to rule out an adverse consequence
of even 1 in 2000.’

Despite his own trial, which is the largest ever carried out, he says
this still holds true. ‘I’m not prepared to say ultrasound is not safe,’
he says. ‘But in the context of doing a test which we have proven definitively
has no benefit, any potential concerns about safety have to be considered.’

Bucher and Fraser say a trial designed to consider all possible adverse
outcomes should be carried out. But such a trial would not be easy to arrange
now that most women are scanned.

So why has such a trial never been done? Nielson says the rapid development
of technology meant there was a danger that by the time a trial was completed,
the imaging technology would have changed. In 1976, the Medical Research
Council (MRC) set up a pilot trial, and then abandoned it because it was
thought unlikely to reach definite conclusions. Over the next decade, ultrasound
became routine in many British hospitals.

Conflicting conclusions

In 1985, the MRC hosted a meeting of obstetricians, epidemiologists,
biologists and medical physicists to discuss screening. While there were
good reasons for expecting routine scans to be beneficial, the meeting concluded
that: ‘No definitive data on the benefit of routine scans was available.’
It added: ‘Epidemiological data on the risks associated with antenatal ultrasound
scanning were inconclusive.’ It recommended that a working party should
be set up to advise the council on which questions needed to be answered
and the most effective design for a study. The working party was never set
up.

In 1984, a working party of the Royal College of Obstetricians also
called for a study to compare routine with selective scanning. This study
was never carried out.

In the US, in the same year, a consensus conference organised by the
National Institutes of Health, considered the same evidence as the college
and ruled out routine scanning. The procedure is still not routine in the
US, although many women have a scan on the advice of doctors.

Fraser says the research called for in Brit-ain was not carried out
because doctors felt ultrasound was so likely to be safe it was unjustified.
Ewigman sees this attitude as part of a pattern. ‘We see this over and over.
Fetal heart rate monitoring was the same,’ he says. ‘It is quite difficult
at least in the US to get drugs approved. But we don’t have the same approach
to technology.’

Ewigman quotes a model for technological development which states that
promising early reports on an innovation are picked up by doctors, latched
onto by the public and adopted as standard before trials are carried out.
Critical evaluation comes later. Doctors, he says, have a desire for diagnostic
certainty and tend to accept early research findings uncritically.

For all the debate over the clinical value and safety of ultrasound,
the decisive pressure for a re-evaluation may come from another direction
– economics. ‘Increasingly, everything has to be evaluated from a cost-effectiveness
point of view,’ says Benjamin Sachs, chief obstetrician at Harvard Medical
School. The pressure to reduce costs may force doctors to be more critical,
he says. It may also settle the doubts over ultrasound once and for all.

Helen Saul is a medical writer based in London.

Joints contain sensory nerve endings called proprioceptors, which normally
send signals to the brain and to the surrounding muscle via the spinal cord.
These signals tell the brain what position the joint is in and also influence
the activity of the surrounding muscle.

William Ferrell of the Institute of Physiology at the University of
Glasgow says that the inflammation that occurs in rheumatoid arthritis may
damage the proprioceptors, causing them to send false messages to the brain
about the position of the affected joints.

He came to these conclusions from experiments involving people with
rheumatoid arthritis in which their index fingers were moved without their
being able to see them. They were then asked to move computer-controlled
facsimiles of the fingers until they matched their new position, still without
looking at their hands. A control group, matched for age and sex, were also
asked to do the same task.

Ferrell and his colleagues found that people with rheumatoid arthritis
have a ‘flexion bias’: they tend to think their fingers are straight when
really they are slightly bent. He believes the damaged proprioceptors may
be sending false messages to the brain, causing the flexors to work harder
than they ought to. Eventually this produces the permanently flexed fingers
that are characteristic of rheumatoid arthritis.

According to Ferrell, standard physiotherapy, which aims to build up
the flexor muscles, could exaggerate this tendency, and better understanding
of the condition could help physiotherapists devise more rational regimes.
‘Patients with arthritis would undoubtedly become worse without the only
currently available treatment. But if in future we add other regimes to
the existing ones, we may be able to minimise deformities.’

Exercising the extensor muscles might help to counteract flexion, minimise
deformity and improve function at affected joints, says Ferrell. He stresses
that his work is still in its early stages, but suggests that in future
electrodes stuck onto the backs of the hand might provide relief for people
with arthritis by stimulating extensor muscles through the skin.