¿ìè¶ÌÊÓƵ

More than four years after it was first licensed, the AIDS drug zidovudine
(AZT) is still on trial. Many thousands of people in the US and Eu-rope
now take it, but there are serious unanswered questions about its long-term
effects.

When zidovudine received its licence in 1987, it was intended for treating
people with AIDS. Since 1989 in the US and 1990 in Britain, it has also
been approved for people who have HIV infection but no symptoms of disease.
It is the use of the drug for this second, much larger, group that has fuelled
controversy. While there is clear evidence that the drug prolongs life for
people with AIDS, the evidence in its favour for people without symptoms
is much less clear.

Trials have shown that the drug delays the progression of disease in
some symptom-free people. But none of the studies has lasted long enough
to determine whether giving the drug at this early stage actually prolongs
lives. Nor have they continued for long enough to tell us about zidovudine’s
long-term effects.

Despite these questions, many doctors now prescribe the drug early in
the course of infection, believing this to be the best action. But some
researchers – and some patients – are not so sure. Not only does zidovudine
have unpleasant side effects, it may also lose its power against the virus
over time. After a period of treatment, strains of HIV that are resistant
to zidovudine tend to appear in a person’s blood.

It was to resolve these issues that Britain and France set up a joint
study, called Concorde, a placebo-controlled trial of zidovudine in HIV-positive
people without symptoms. The study began in 1988 and earlier this month
an independent group of assessors decided it should continue.

So far, the most authoritative evidence of zidovudine’s effects in people
with HIV but no symptoms, comes from a large-scale trial organised by the
US government’s AIDS Clinical Trials Group, which was stopped early and
its results announced in an unprecedented manner, via a press release (This
Week, 26 August 1989).

This trial, popularly known by its protocol number ‘019’, had found
clear evidence that the drug delayed the progression of disease in symptom-free
people whose immune systems had deteriorated beyond a specific point. This
point is measured in terms of the number of CD4 cells – the type of T cell
that HIV attacks – in the patient’s blood. Healthy people have about 1200
CD4 cells per cubic millimetre of blood. People who are beginning to develop
AIDS usually have a CD4 count below 200.

Patients in 019 had been followed up for an average of 55 weeks when
an independent board of monitors decided the trial should stop. The 1338
participants had been divided into three approximately equal groups and
given a high dose of zidovudine, a low dose, or a placebo. In the placebo
group, 33 had developed AIDS, compared with 11 in the low-dose group and
14 of those receiving the high dose. The monitors decided that because of
this disparity it would be unethical for the trial to continue.

Investigators involved in 019 found that zidovudine delayed disease
in symptom-free people with a CD4 count below 500. All of those with counts
below 500 were offered zidovudine as soon as the trial ended, even if they
had previously been on placebo. For people with a higher count, the results
were not clear. The only people still in a placebo-controlled trial of zidovudine
in the US are about 1000 participants of 019 whose CD4 counts are still
above 500.

Concorde, on the other hand, was kept running after the American announcement
and will now carry on for at least another nine months, if not longer. The
decision to continue was made by an independent monitoring committee similar
to that which called a halt to 019 in the US. These committees are allowed
to ‘see’ a trial’s interim results, unlike the clinicians and patients involved,
who remain ‘blind’ to who receives zidovudine and who gets placebo.

In its statement this month announcing the continuation of the trial,
Britain’s Medical Research Council said that the monitoring committee had
found ‘no clear evidence on which to base new recommendations for clinical
practice’. But, it continued, the committee considered that ‘if patients
remained in the trial for a further possibly limited period, additional
information on important issues including the longer term risks and benefits
would be obtained. The (committee) felt ‘justified, and indeed obliged’,
to recommend continuation of the trial’.

Concorde investigators assume that the trial would only have been stopped
if it had shown a clear and significant benefit or harm that had not shown
up in 019; for example, if there were unexpected deaths in the placebo group.

The reason why the US trial stopped and the European study continues,
lies partly in the stark social differences between the two sides of the
Atlantic. The sheer scale of the US epidemic has raised a correspondingly
louder call there for a rapid solution – any solution. In the US, about
1 million people are infected with HIV and about 200 000 have AIDS. Britain
and France, although seriously affected by AIDS, have far smaller epidemics.
In Britain, there have been just fewer than 5000 AIDS cases.

In addition, the US has no national health service. According to Nick
Partridge at the Terrence Higgins Trust, this increases the push to get
AIDS drugs licensed quickly. Normally in Western countries, the only way
a person can receive an unlicensed drug is to participate in a trial of
that drug. In Britain and France the health services, within which such
trials are done, reach to even poor inner-city areas. But in the US people
in the poor inner-cities rarely have good enough contacts, or access to
the right clinics, to gain acceptance to trials. The pressure for drugs
to be licensed – and so available to all – is consequently greater. ‘People
with AIDS and HIV in the UK are in a much better position,’ says Partridge.

But the desperate need for a solution has to be balanced against the
need to confirm the benefits of treatments in properly conducted scientific
trials. ‘Because of the social pressures of this disease, there is pressure
on us as clinicians to close chapters early,’ says Ian Weller, professor
of genito-urinary medicine at the University College and Middlesex Hospital
School of Medicine in London. Weller is chairman of the Concorde working
party.

The original objective set for Concorde by its organisers, the MRC and
the French equivalent, INSERM, was to assess over at least three years whether
zidovudine given to people at an early stage in infection both delayed disease
progression and prolonged life. So far, participants have been followed,
on average, for 20 months, so time is not yet up. The decision to continue
it now has not surprised those involved.

It is important to stress, though, that no British or French patients
who need zido-vudine are being forced to risk receiving a placebo. Since
the Americans broke the code on 019 and offered the drug to everyone with
a CD4 count below 500, the Anglo-French trial has offered patients the chance
to ‘opt out’ and receive zidovudine, with their doctor’s agreement, once
their CD4 count has fallen below 500 and is persistently declining.

Among the long-term effects that clinicians hope Concorde will shed
more light on, is the unpleasant anaemia that zidovudine causes. Studies
reported at the 1991 international AIDS conference in Florence, suggested
that only a small percentage of users have severe anaemia or other toxic
effects. Some researchers, such as Paul Volberding, the leading investigator
on 019 at San Francisco General Hospital and a man with long experience
of treating AIDS, says he doubts that longer-term studies will produce any
surprises about these effects.

But almost everyone – including Paul Fiddian, head of clinical virology
for Wellcome, which manufactures zidovudine – accepts that the question
of drug-resistant strains needs much more study.

Zidovudine, is known as a nucleoside analogue, as is its sister compound
ddI or didanosine which was licensed this month in the US. Zidovudine’s
chemical structure is similar to that of thymidine – one of the four chemical
bases or nucleosides that form the building blocks of DNA. When HIV infects
a cell, it can only replicate by converting its genetic material, RNA, into
DNA with the help of its enzyme reverse transcriptase. Zidovudine blocks
the building of the virus’s DNA by slotting itself into the chain of nucleosides
in place of thymidine. In the laboratory, researchers have identified five
mutations in HIV’s reverse transcriptase gene that make it resist zidovudine.
However, no one knows how much difference these resistant strains make in
the living human body. It seems likely that they will have an important
effect on treatment, but this question must still be answered.

Douglas Richman, professor of pathology and medicine at the University
of California, San Diego, says small studies suggest that resistant strains
are slower to emerge in people with higher CD4 counts. In essence, the sicker
the person, the more rapidly resistant strains appear. For example, many
people with AIDS acquire resistant strains within as little as six months
of starting to take zidovudine. A person with early symptoms of disease
can wait 18 months before resistance appears. No one knows why this happens.

The 019 trial, as its investigators acknowledged, revealed nothing about
the problem of resistance. Indeed, some researchers believe the decision
to halt the study early, limited the value of many of its findings.

‘What does 019 tell us in the long term? Not a lot,’ says Weller. In
his view, evidence from the trial, as far as it goes, suggests that a relatively
small number of people are going to benefit from zidovudine in the short
term. Against this must be weighed the uncertain outcome – such as toxic
effects and possible generation of resistant virus – for the majority.

Weller and others appreciate the pressure that the US researchers were
under, but want more information. ‘The philosophy of Concorde is that drawing
a line at 500 CD4 cells is artificial,’ says Fiddian. ‘So rather than do
what 019 did, they (Concorde investigators) have tried to keep the whole
thing together.’

But some see limitations in Concorde. For example, Volberding in San
Francisco says he wonders whether Concorde will be able to shed much light
on survival times, because people who understandably opt out of the trial
and start taking zidovudine are lost from the statistical analysis. He also
doubts whether the extra months the drug may give are as important as the
improved quality of life it can confer on patients.

Janet Darbyshire, the MRC’s principal coordinator for the trial at the
Royal Brompton Hospital in London, believes that the numbers opting out
of Concorde should not prejudice its analysis too severely. Many people
wait, she says, until their CD4 counts are considerably lower than 500,
probably at around 300, before they opt for zidovudine.

Volberding believes there is mounting, if indirect, evidence to show
that zidovudine prolongs life and improves the quality of life. A recent
study in Maryland compared survival times for people with AIDS diagnosed
between 1983 and 1989. On average, people diagnosed between 1987 and 1989
lived 140 days longer than those diagnosed before and during 1985. Zidovudine
was licensed in 1987. Those who received zidovudine survived on average
for 770 days, compared with 190 days for those who did not. Of all factors
influencing survival, zidovudine was the strongest, said the researchers
(New England Journal of Medicine, vol 324 p 1412).

Yet it was not the only factor. Since the mid 1980s, doctors have got
much better at detecting and treating the secondary infections that exploit
HIV’s destruction of the immune system. It is these infections, such as
Pneumocystis carinii pneumonia and toxoplasmosis of the brain, that actually
kill people with AIDS. The medical improvements may account for an important
part of the increase in survival time.

Hopes that the development of new generations of HIV drugs will make
arguments over zidovudine redundant, are still unfounded. The next arrivals
are some years away. In the meantime, researchers are looking increasingly
at combined therapy with zidovudine and ddI. Used either at the same time
or one after another, they may increase the ‘window’ between starting treatment
and the appearance of drug-resistant strains. The importance of this strategy
was brought home recently with the discovery of the first HIV strain to
show resistance to ddI (Science, 27 September, p 1557). Other, as yet unlicensed
nucleoside analogues such as ddC are also being studied as part of combination
with therapies.

So, the debate over zidovudine continues. Weller sees the sense in stopping
the virus replicating as early as possible. ‘But I’m not sure if we’ve got
the right tool at the moment,’ he says. Some people with HIV and their advocates
agree: ‘I certainly could not say definitely that ‘AZT is a good thing for
you’,’ says Partridge. Volberding, on the other hand, says the effect of
the drug is clear and believes that early treatment is the best course of
action.

Next June, Concorde’s independent committee will analyse the data again.
Until then at least, the jury is still out.