Martin
Fisher, Consultant in HIV Medicine, Royal Sussex County Hospital, Brighton
Margaret
Johnson, Director of Royal Free Centre for HIV Medicine, Royal Free Hospital,
London
Anton
Pozniak, Consultant in HIV Medicine, Chelsea and Westminster Hospital, London
Caroline
Sabin, Statistician, Royal Free and University College Medical School, London
Individual abstract numbers link to the abstracts online at http://www.retroconference.org.
Where session titles are underlined, this link takes you to the list of all abstracts presented in that session.
Where presenter names are underlined, this presentation can be viewed online, and the link will take you to the list of sessions which can be viewed online.
Structured treatment interruptions
KA: The major session today covered
structured treatment interruptions. Did you feel the studies [oral session: abstracts 288-293
and poster session abstracts 354-365] clarified what is meant by an STI and when it might be useful?
AP: Although Bruce Walker said there’s one
type in acute infection and one in chronic infection I think there is one where
treatment has commenced in hyper-acute infection, one in so-called acute
infection and one in chronic infection.
KA: What’s the definition of hyper-acute
infection?
AP: When you’ve got someone before they
have actually converted to antibody. That’s where there appears to be the
highest chance of success.
MJ: So that was the very practical one,
where you had to catch them before they have fully seroconverted, which must
apply to what percentage?
AP: Very small. But it’s practical when
you’ve got somebody like Eric Rosenberg who really looks at the patients to
identify such cases.
MF: Even then it depends on which antibody
test you use as to how early you are going to pick them up. So, just like we
are saying we are going to need a definition of STIs, I think we are going to
have to redefine primary infection to answer that.
MF: Yes, in terms of STIs there was the
hyper-acute versus acute and the acute versus chronic. There is also the [sub-division between]
during-virologic-success-having-a-break, or the treatment interruption during
virologic failure. The trouble is they were all hotch-potched together into one
amalgam when they are looking at very different issues.
AP: - and then there’s treatment
interruptions to try and prevent toxicity where you just say ‘let’s stop
because you are being so ill on all these pills’
KA: - which is probably the most common
reason for a treatment interruption at present -
AP - which they didn’t actually address at
all today.
MF: There were no presentations that looked
at whether intermittent therapy altered toxicity profile. Nothing.
KA: What do you mean by intermittent
therapy?
MF: Either pulsed therapy, as in the two
Fauci posters whose aims as presented at Durban were to reduce toxicity or to
reduce costs, yet it it struck me that there was no toxicity analysis and that
the level of virological and immunological follow-up that they were advising
might actually increase costs rather than reduce them.
CS: I also think, especially the short ones
that Fauci is recommending, which is seven days on, seven days off, on top of
the cost, I think it’s also going to reduce quality of life. I really feel that
having to worry about what day of the week is it, what week is it, whether you
should be on or off, actually may make things a hell of a lot worse.
AP: I strongly believe it’s not an issue
for the developing world. To say that it might be useful for the developing
world, without actually building any capacity and infrastructure in the
developing world, is going to [make it very difficult to implement]. And I
agree with Martin, that it might cost a fortune to make sure people are doing
it all right and monitoring it all.
MJ: And I think when you talk about
‘quality of life’, the other thing is adherence. In the clinic, one thing that
I notice is that patients often, if they feel they have lots of toxicities and
it is interfering with quality of life, want to stop therapy. And very often it
focusses their minds as to why they feel better. No one has really addressed
what then happens for that group of patients when they go back on therapy.
CS: Well I think that is probably almost
the same group as patients with virological failure in the observational
databases because those people are choosing to stop therapy themselves, for
whatever reason, and probably a lot of those reasons could well be toxicity, or
quality of life issues. So I think we’ve got some information from
observational databases but we still don’t know, from any of those studies,
whether those treatment interruptions are actually beneficial, whether they are
harmful or whether they have no impact at all. We are really lacking that data.
MF: There was a presentation (abstract 292)
on using treatment interruptions in patients in so-called salvage therapy.
First of all - no definition of salvage. The patients actually had another
option – half of them did – they had NNRTIs to go to, and some of them were
given T-20. The bottom line for me was that he didn’t show that either just
switching the patients onto a new therapy or doing nothing, would
have ended up with the same situation. He tried to make out that having the
drug holiday improved the outcome in those people with one drug class left to
use, but actually he didn’t do that at all because there was no control to show
that that was the case.
CS: That’s actually a problem which was
common to almost all of the presentations we saw in that session. In very few of
them was there any form of control group. In the one presentation where there
was a control group, it was a silly control group to have anyway [abstract
289].
MF: They had been randomised 3 to 4 years
earlier so they had actually ceased to become a control group by the time the
STI element came into the study.
CS: So it was patients who hadn’t received
treatment. So they were comparing patients who had taken treatment for a while
and then undergone a treatment interruption compared to patients who hadn’t
ever received treatment rather than the probably more relevant comparison of
people who had started therapy at the same time and stayed on it. So there was
a lot of debate in the audience about the value of it and whether the treatment
interruption had any benefit whatsoever.
MF: One concern that I have is that one of
the main benefits everybody appears to be aiming for is improved immunologic
control yet the one consistent finding in every STI study is that your CD4
count drops, or at best, stays where it is. So, arguably, whilst with some of
the very early infection studies the viral load comes down, there seems to be
no data to suggest any immunologic benefit that may have clinical relevance
yet.
KA: And if there is any contribution from
HIV-specific immunity to immunologic control of HIV, it’s certainly not being
reflected in those CD4 changes. And in all the other studies, those people are
very much a minority.
AP: In the Deeks study [abstract 292], 3
people got an opportunistic infection during the STI. Now I don’t know whether
they would have done in any case if they had stayed on HAART, but that was a
worry to me. And then some other people went onto get non-Hodgkins lymphoma and
other things after resuming therapy. The CD4 drop was 88 cells and the viral
load rose by 0.74 log. Now, you might be OK if you did a controlled trial with
people who had CD4 counts at 600 because if you lose 88 cells its not so
disastrous but what about all those people who are grabbing STIs at 200?
style='mso-bidi-font-style:normal'>
MJ lang=EN-US>: You probably wouldn’t be on treatment at 600
CS: In the work we’ve been carrying out with
Veronica Miller (abstract 365), the people who start off with 600 CD4 cells are the ones who
have the biggest drops.
MF: There’s a poster here to show that
people who had the biggest drops had had the biggest rises in CD4 count when
they started treatment.
CS: So unfortunately, whatever your nadir
was before treatment, the moment you stop treatment, your CD4 count plummets
back down towards that level.
AP: Which just goes to show that thymus
regeneration might take a long time.
MF: Similarly, [looking at] the viral load data, in one
of the presentations, the people with the highest viral loads were the people
who rebounded the quickest and to the highest [level]. Which all seems to point to the
fact that when you interrupt therapy you go back to where you were before,
which for some people may be fine if they started treatment very early. But for
people who started when they were sick, which as we discussed yesterday is a
significant number of people in the UK, then the risk of having an STI may well
outweigh any theoretical advantages.
AP: The bottom line is that if you are
going to have an STI, go into a study.
All: But there aren’t any.
AP: Well, you really need to be monitored
very carefully if you decide you are going to have one.
MF: the good bit of news I thought was that
every study that looked at it showed no evidence of development of resistance.
AP: But they didn’t look at minority
species. You may accumulate minority species over time – one mutation here, one
mutation there – and we don’t know whether that happens because the tests
they’re using are too crude. There’s a very good paper by a guy called Hance (abstract 293
showing they can detect minority species in people who have reverted to
so-called wild type during a STI
KA: What’s the sensitivity, what’s the sort
of minority that is detectable?
AP: 1% for people with L90 and 1% for
indinavir mutations.
MJ: There is one case of resistance, in the
Fauci study, in the one third of time off, two thirds on treatment. One patient
who was on quadruple therapy, who perhaps might have had some nucleoside
treatment previously - they don’t characterise that patient –
developed NNRTI resistance both genotypically and phenotypically.
CS: I think there’s a real need for a
randomised trial. As a statistician, I’m a great fan of observational databases
but we are having terrible problems trying to work out whether treatment
interruptions are good or bad. I don’t believe we will ever really know the
answer from observational data. We
really do need a trial.
MF: It comes back down again to doctor-
patient communication, as well, and the need for openness and discussion, to
discuss the risks as well as the benefits and do it in partnership rather than
in isolation without being aware of the potential down side.
Transmission and primary infection
MF: I have to say that what I thought was
interesting was in the study on gay men in Amsterdam (abstract 261), where the
frequency with which people engaged in unsafe sex was related to their last
surrogate marker responses and it appeared that, behaviourally, people were
getting the wrong message from an undetectable viral load - that they were
unable to transmit - which is clearly not the case as today’s poster session
showed, in that a quarter of people with undetectable viral load have detectable
virus in another compartment, especially the male genital tract, which may be
transmissable.
AP: I think the most important thing from
the Rakai study which looked at the probability-per-act of transmission in
HIV-discordant couples in Uganda(abstract 266) was that if you were a
circumcised, HIV-negative man with an HIV-positive partner you didn’t get HIV
from your positive partner if you were circumcised. That was quite interesting.
There was no risk of acquisition in that study.
CS: The interesting thing I found in that
study is that they found a very strong age effect, with the probability of
transmission declining with age.It wasn’t just a function of the fact that
younger people tended to have more sex, it was actually per act so there is something
extra going on. One of things we were discussing at the time was whether,
perhaps, there is a bias there to do with the way people discuss their sexual
behaviour and maybe in younger couples they may be under-reporting or
over-reporting sexual acts.
AP: It may also be physiological. In
younger people the vagina or penis may not be so keratinised or intercourse may
be more traumatic.
MF: Or it could be that, given that there
was also an association in that study with herpes simplex infection, the natural
history with herpes simplex is that episodes get less frequent as one gets
older and that may be what is causing the relationship with younger age and HIV
transmission.
KA: Certainly in terms of acquisition in
women, younger age is shown to be very strongly correlated with the risk of
infection, and more specifically, age of sexual debut in women, because the
cervix is immature and more vulnerable to infection.
AP: There was a very interesting paper from
the CDC, Atlanta arm on the prevalence of mutations associated with decreased
antiviral drug susceptibility amongst recent and chronically HIV-infected
persons in 10 US cities from 1997-1999 (abstract 265). What they showed was
that the rate of NNRTI resistance - primary or secondary mutations - was 1.7%,
NRTIs 6.3% and PIs 1.2% and they found that there was higher prevalence in men
who have sex with men, Caucasians and partners of people who are on
antiretrovirals.
CS: Presumably the first two just related
to access, and they are more likely to have partners who are being treated as
well.
AP: Now, I was going to ask: if you had a
partner who was being treated and you had unsafe sex, would you take their
pills? That may be a factor, and then you get resistance because you’ve taken a
few pills but not very many.
KA: Do you think that knowledge about post
exposure prophylaxis is so widespread that people would actually be doing that
kind of ad hoc PEP?
CS: In some groups of people I should
think.
AP: I thought this was also very
interesting, that there was a variation geographically through the United
States. He said it was quite marked, although he didn’t give all the figures.
CS: There was a distinction between the
Western USA and the non-western area.....
MF: The UK data, which has been generated
through the MRC seroconverters study and Deenan Pillay’s lab in Birmingham,
suggests that there isn’t any variation within participating centres in the UK
.
MF: On the subject of primary infection
there was also a poster suggesting quite a high level of sexual transmission
preceding symptomatic infection (abstract 411). Traditionally, one of the
reasons for diagnosing primary infection is to reduce transmission but actually a
lot of the transmission is happening in the few days leading up to symptomatic
illness, but the peak of viremia is before symptoms occur. This peak occurs as
early as day five after infection.
KA: There has been epidemiological
modelling which has looked at the contribution of high rates of partner change
amongst people in primary infection which has suggested that this is the
primary motor of ongoing transmission amongst populations where there is
already a high prevalence. So it is not altogether surprising that they’re
detecting those levels of viraemia and cases of ongoing transmission.
CS: I guess the point that you’re raising
though is that trying to catch people and treat them is not going to have an
impact on the majority of those transmissions.
MF: While it may have an impact on
individuals, it may be naive of us to expect that it is going to have an impact
on populations.
CS: We pick up so few of those people
anyway at primary infection. And if you’ve missed the peak of viraemia because
it’s just happened a few days prior to that, then you may really not prevent
many of the new infections.
MF: I don’t know to what extent we are
missing it. I think we may be misdiagnosing it, and I think it may be
presenting to the wrong care provider who is lacking appropriate skills to
diagnose it. I think it may well be presenting to the healthcare settings but
it’s missed.
CS: So for example, GPs are not looking for
HIV in a person who they don’t perceive to be at high risk.
MF: And the current diagnostic tests we use
in the UK are not sufficiently clever to pick up recent infections.
When to start revisited
AP: For the BHIVA guidelines, I think
people should look at abstract 341, Long term survival after initiation of
antiretroviral therapy, which again shows no difference in the risk of death
between those starting above 350 and those starting between 201 and 350 CD4.
MF: In fact, there are two abstracts side-by side
both suggest that if you start therapy with a CD4 count of less than 200, then
you have an increased mortality [see also abstract 342].
AP: But over 200, there’s no difference in
this cohort of 201-to-350, or great than 350 – the Kaplan-Meier survival curves
were the same. But once you get below the 200 threshold, Julio Montaner’s shown
that, Andrew Phillips has shown that and now all these other studies are coming
through. Andrew Phillips is looking
at when to start in terms of surrogate markers responses to treatment. Julio
Montaner’s is in terms of survival
CS: Could I just intercede with something
which isn’t actually at this conference but which will be presented soon, which
is a huge meta-analysis of patients starting therapy and looking at their
responses according to their baseline CD4 count. And that’s going to be
presented at a meeting in Monte Carlo in the next couple of months at the
Cohorts meeting, which is pulling together data from all of these reports.
Adherence
KA: Have we learnt anything new about
adherence at this meeting?.
MF: What struck me with adherence was quite
a big paradigm shift away from trying to label which groups of people are
adherent or non-adherent, which clearly doesn’t work and doesn’t particularly
help people. And for the first time in HIV there were three randomised
studies looking at different interventions (abstracts 479, 480 and 481). One intervention
being an education program or individualised social support, one being an
antiretroviral information clinic with subsequent individualisation of therapy
based on lifestyle, and one being an internet alarm clock that would tell you
when to take your medication (abstract 480).
KA: It sent messages to people’s pagers.
People were actually given a pager by the clinic and the effect was compared
with individual counselling.
MF: Yes. Each of those studies suggested an
improvement in adherence. Though in two of those studies they followed
adherence longer term and it suggested that in the behavioural interventions or
the education programmes, the effect was lost by 6 months. What I think it
tells me is two things: one is that very different ways of tackling adherence
may all have an additive benefit; and secondly, that it’s no use just producing
that in the short-term, but that it needs to be incorporated as part of a
longer term adherence support program, rather than a one off programme for
people starting therapy.
CS: Did any of those studies take it one
step further and look at surrogate markers and response?
MF: Yes, one did
CS: And did that translate into benefit?
MF: Yes it did. They looked at the
relationship between adherence and time to development of an AIDS-defining
illness and found that every 10% difference in adherence rates was associated
with a 21% reduction in the risk of illness over 28 months of follow-up
[abstract 483], which suggests to me that spending money on adherence support
is highly cost effective.
Hepatitis B and C
AP: There was just one thing today I saw
today. Interferon, ribavirin and the use of ribavirin to boost ddI (abstract 308.
KA: What did you think about that?
AP: Well I think it’s interesting. You
might think if you are going to treat people with hepatitis C and HIV disease,
you might put ribavirin into the formula [where it is synergistic rather than antagonistic with anti-HIV drugs].
KA: Is there a case, given we
don’t really know whether to treat for hep C first or HIV first, to use ddI as
a sort of boosted monotherapy for HIV, given the antagonism between other NRTIs
and ribavirin [and the much higher IC90 achieved when it's co-dosed with ribavirn and alpha interferon], whilst using ribavirin plus interferon to initially deal with
hepatitis C.
AP: That’s an interesting concept but the
data from the hep C cohort studies suggests that you may need to control your
HIV disease first.
KA: - but Caroline is very keen to have a
randomised trial on this question!
AP: You might need to do that. There is
some data to suggest that that is probably the thiing to do – control your HIV
disease and then treat the hep C. Because if you try and treat hepatitis C with
uncontrolled advanced HIV disease, you have a reduced chance of clearing HCV.
style='mso-bidi-font-style:normal'>
CS: The only concern is that, as we were
saying yesterday, is that because the HIV treatments [are] such long-term, life-long
things, you are going to be giving people three drugs, say, and then adding in the
ribavirin and interferon on top of that.
AP: I think that that’s OK if you stabilise
yourself on the HAART regimen and you are used to it and your lifestyle is
fine, then you can start doing that. The thing that might put me off ever going
on HAART might be a terrible experience on interferon/ribavirin, leading me to
think that all this treatment for my disease is all completely hopeless. That’s
just a thought, not based on any data whatsoever!
Efavirenz versus nevirapine
KA: There
was some provocative data from Andrew Phillips comparing clinical outcomes on
efavirenz or nevirapine-containing HAART in the EuroSIDA cohort [(abstract 324].
What did you think of it?
MF: The EuroSIDA data suggested that in
a population which was very largely
treatment-experienced (about 95% had prior nucleoside and/or protease
experience), that when other factors were controlled for, efavirenz appeared to
out-perform nevirapine.
As Professor Phillips quite wisely pointed
out, it was observational data and one needs randomised trials. Nonetheless, it
was provocative data. My worry would be that this isn’t the population in which
non-nucleosides have been widely used in the UK, and given style="mso-spacerun: yes">
naive individuals, we may have to wait for prospective trials to draw such
conclusions.
AP: And I wonder whether the nucleoside use
was controlled for, because a lot of people used to give nevirapine, ddI & d4T
which was more difficult to take because of the adherence on the old ddI, and a
lot of people gave efavirenz/Combivir because of the history of how it was
studied in DMP-006 and other trials. So what I would say to Andrew Phillips
would be: have a look at the database and match up the d4T/ddI with the AZT/3TC
– because you may not find a difference then.
CS: I think we
should watch this space because I’m sure there are studies out there which are
now trying to look at this in naive patients [Editorial note: the 2NN study is
currently comparing 2 NRTIs plus nevirapine alone, efavirenz alone or efavirenz
plus nevirapine in treatment naïve patients]. Again, one of the problems with
observational data is that you really have to be careful about the
interpretation. If we see the same trend appearing in three or four databases
then maybe there really is something in it, it may not simply be a function of
bias.