Salvage Therapy

Dr Mike Youle

This is a transcription of the training session salvage therapy and HIV that was given to the UK-CAB by Dr Mike Youle. The full set of 84 slides to accompany this talk are also available on the i-Base website with the other reading material for the August 2002 meeting.

The talk provides an overview or the range of treatment approaches for treatment people who have developed resistance to current HIV drugs together with ways of making any HIV treatment more effective. Transcription edited by Simon Collins.

I’ll start by saying that there are:

  • lots of bad doctors,
  • lots of bad nurses,
  • lots of bad patients and
  • lots of bad drugs!

A lot of these slides were part of a talk I gave about six months ago and most of it is still relevant – there is also plenty of time for questions.

Setting the scene: as time goes on you get better at everything.

Just as I am much better at writing, and many other things, than I was when I was twelve I think we are much better at looking after people with HIV now than we were a decade ago. That is because of a number of improvements, mainly in the options for treatments that we have got but also because of experience. What I mean about bad doctors and nurses really refers to their lack of experience.

Bad patients and bad drugs are a different issue that we can discuss in a minute, but bad doctors and nurses basically means inexperience, or people who think they know more than they do, or people who are just not interested in finding out the latest information. The major determinant of whether you are successful with your therapy is a combination of reversing those four issues above rather than new science or research.

We also have improvements in the availability of things that are outside of the immediate health sphere. What somebody does at Imperial College, or Roche Pharmaceuticals, will affect what we in the clinics can do to make your health better. That is a continuous process that has been speeded up dramatically from what it was a few years ago – not necessarily just because people wanted this but because we now have new technologies.

When I first started trying to get a new drug tested to see if it was effective against HIV it was very difficult. When I first got involved in this at all HIV wasn’t even identified so you can’t test a drug against something when you don’t know what it is. With an identified virus you first have to work out what it is, then you’ve got to work out how to affect it, but you might not have the technologies to test drugs with it. So even in 1986/87, when we were testing AZT and ddI, it was not possible to effectively evaluate how good that drug was because the technology was not there.

Now we have screening processes that allow you to evaluate carefully, every day, what hundreds and hundreds of compounds might do in the test tube. Economies of scale have improved things.

And with our therapeutic options we are keeping pace with clinical disease – but only just.

I would argue in this country with good treatment, 99.9% of patients with HIV could have a reasonably healthy life compared to a few years ago. However we do have rising rates of toxicity and we do have increasing resistance. We also have increasing three-class experience to nucleosides, non-nukes and PIs. Whether you define nucleotide inhibitors (tenofovir) as a new class is another matter but we certainly now have fusion inhibitors and receptor blockers are also going to come through soon. This is perhaps the most exciting development from the last 6-12 months.

There are also new competitor diseases. I went to a lecture that reported nineteen new infections in humans, in one year, that hadn’t been seen before. That happens year on year – so HIV is just another one of these. This week at the Royal Free we’ve had two new sexually transmitted hepatitis C infections and at Chelsea and Westminster and UCH they are now seeing something similar which was supposed not to happen, i.e. sexual transmission of hepatitis C.

We are now beginning to screen for HHV-8 which is the virus linked to KS (Kaposi Sarcoma), and in addition to new infections a lot of the old ones like syphilis continue to cause trouble if you don’t keep an eye on them.

In 1997 the Palella data (slide 3) showed that HIV therapy – and not just PIs as shown in this slide – lead to a huge reduction in deaths. But over time and with increased use of therapy, people also become better at providing therapy. The benefit is partly attributed to this take up of therapy and partly to improving experience of the doctors and nurses and other health care workers.

And year on year exposure to different drugs increases. At the Royal Free in our 1500 patients, on HAART, we are doing quite well. In 1999/200, about 82% of viral load results are less than 400. Updated data for this year shows slightly better surrogate marker effect i.e the low viral load and higher CD4 that seems to lead to longer and better health quality. However already 30 % of our patients have all three classes and of those 26 % have effectively failed – I don’t really like that word because I don’t think it necessarily conveys what it means – but what Clive Loveday was saying 26% have got greater than 400 copies so perhaps they are more likely to have progression in the future. However on the good side only 6 % have got less than 50 CD4s which we now know is a very strong marker of continued risk of health.

My comment about disease becoming more varied, and the first remark about patients are linked by (what seems like) every gay man in London who is barebacking – or having unprotected sex. Either way you run the risk of getting multidrug resistant virus and hepatitis C. I did this slide two years ago and I have not changed it. For gay monopoly, you do not pass go, do not gain any or much benefit coming from treatment – so you are going right back to where we were 15 years ago before antiretroviral therapy existed.

What is the goal of therapy?

The goal in science may be the only acceptable viral load being the one that is undetectable using the most sensitive assay – and our tests now go down to less than 50 copies. Hopefully Clive is going to be doing a less than three copy test soon and there is some quite exciting data from the Seville Resistance Workshop a few weeks ago about this. Giving another HIV drug to people who were stable on therapy, with between 10 and 30 copies, made their viral load goes down another log (to <3 copies). We don’t know what that means in terms of improved health or lower risk of rebound and lower risk of resistance, but what we do know is that you can achieve it and that every log drop in viral load has been seen previously to be beneficial.

There is a good study from the Frankfurt cohort showing your risk of failure and your risk of resistance goes down with every log you drop. So it may be that adding in a new drug may make a difference. A study at the Free will randomise people to add in tenofovir or placebo when they’ve got a viral load less than 50, to see if forcing the viral load down any lower might make any difference. We don’t know if it will but I think it is worth doing.

The practical benefit in clinical terms is achieved when disease progression is stopped and this occurs with a viral load less than 5000 – although at this level this is only a very short term effect.

The reality might be a balance against toxicity, so for someone who has used all of their options you may want to prevent progression as much as you can but not cause toxicity. There are interesting studies looking at replication capacity – how good HIV is at reproducing itself – and that viruses under the effect of a drug may still not be particularly fit and might be a disabled virus, and I have a few slides which I will show you at the end. (slides 77-81)

With each individual you need to decide where you are on that continuum. So perhaps salvage therapy might be defined as treatment strategy to reclaim health when initial therapeutic interventions have failed.

What to measure for evidence of benefit in salvage therapy

  • We could use clinical endpoints.

A clinical endpoint is when you get a new or worsening disease, the most extreme being death. But the famous phrase ‘it is time we stopped looking for body bags’ means that there are very few clinical end point studies running now.

The more rare an event occurs the more difficult it is to prove a difference in treatment intervention. If something is only going to happen once every ten years in a group of patients and you try and get rid of that event, you are going to have to have hundreds and hundreds of thousands of people. So as clinical events have become less frequent, it becomes more difficult to prove that something is having an effect.

Two major end point studies are SILCAAT and OPTIMA. SILCAAT is a study looking at IL-2 in about two and a half to three thousand patients followed for six years to get an effect.

The same is true of OPTIMA. This is a randomised study of different ways to treat salvage therapy of 1300 patients who have already experienced three-class failure. We actually found it quite difficult to find those patients in Britain. Now maybe because we are either better at treating people, or that people are less treated i.e. they have had less exposure to therapy and are therefore less likely to not benefit from the newer treatments coming through.

In addition to clinical endpoints you can look at the surrogate endpoints, like changes in CD4 and viral load. Clearly surrogates are protection -the more CD4s you have the more likely you are to benefit from the outcomes – although you might want to look at the quality of those increases. There was an interesting late breaker at Barcelona from the Swiss group that looked at the effect of vaccination in people by their lowest ever CD4 count. For example if everybody in this room had 300 CD4 cells and you vaccinate everybody, the response to those vaccine would be determined by your lowest ever CD4 count, not by your CD4 count when you got the vaccine. An old question which I still think hasn’t been answered is whether your immune system is the same on the way up as it was on the way down – i.e. what is the quality of those T cells?

Q: Are they were trying vaccines on the wrong people. They should be trying it on the people with the high CD4s as they are more capable of producing an immune response…

A: You are more likely to get an effect with a higher CD4 and that is how most therapeutic vaccine studies are moving forward.

  • Quality of life

What you are looking for is quality of life and there are almost no studies that measure quality of life, although some of these bigger ones now do include this.

What is interesting is that you can sometimes get quite good quality of life outcomes when you least expect them. An interlukin-2 study (of this immune stimulant) last year (and IL-2 is not pleasant to take – you feel awful when you are taking it) showed it actually led to an improved quality of life over the length of the study. So ironically even though people felt awful for a few weeks of the year, they felt better overall generally.

Toxicity rates and severity are becoming more important. If you define a target population where you might be three class exposed – this is where we are at the moment, and with fusion inhibitors here we will have four classes and you might have a class leftover – you have to be careful about how you define ‘exposure’. You might be exposed to a number of agents, but not resistant, because you used treatment with an undetectable viral load and switched treatment because of side effects. One week of saquinavir ten years ago is not exposure.

It is important to use the tools like resistance testing to evaluate what happens and there may be an effect of not having a drug over a very long period of time in terms of reversion to wild type virus and therefore resistance becoming less important.

There are exposure issues – if you don’t take the pills they won’t work. Adherence is important. Actually, once they get into a situation where they feel they actually want to take therapy, people who were not adherent in the past may be much more successful than someone who was very adherent before but for whom the drugs didn’t work particularly well. You have to address people’s health beliefs and I think that is the thing we are worst at doing prior to starting therapy.

Pharmacokinetic issues, which we will look at in a minute, look at how well the drug can be delivered to individual areas of the body.

Coinfection and comorbidity – basically what kind of health you have and whether you have other illnesses – for instance we have a number of diabetics with HIV and therefore there are implications for what kind of drugs they can take and how those drugs affect them.

Toxicity issues limits choices for some people, for instance if you have a severe rash on a drug you may be limited in taking that in the future, although I think sometimes we are not very good at making the effort to get over some of those issues. For instance there has just recently been reported a way of avoiding nevirapine rash or desensitising people who have had a severe nevirapine rash – there is always that fear that another one may occur or it might be more severe this time.

Pregnancy is a very difficult issue because you are balancing the health of the mother against the health of the baby. There are so many uncertainties about what individual drugs may do in the foetus.

There is increasing evidence year on year that you CAN get second strains of HIV and anybody that wanders around believing you can’t is going to have to accept that this does occur. What we don’t know is how easily it occurs, how frequently it occurs, and the clinical implications for the individual.

So, if you want to get to full health you might have to work with your therapy and that may not fit in with your life, or it may not be possible and certainly in some settings the drugs may not be available or you may not have enough money to pay for them. So what are you trying to achieve?

PowerPoint Slide
click on slide for a larger view

Figure 1 (slide 10) is a collection of data looking at progression, but it basically it shows that you don’t get progression of disease if you have a high CD4 count (over 300-400) and low viral load (below 1000 or 10,000). Over a year very few people develop serious illnesses and some of these may very well be those associated with having a high CD4 count like lymphomas for which we may have to have other interventions and ironically may be less likely to occur in certain situations where immune suppression occurs. All the yellow dots show that with a viral load greater than a 100,000 and a CD4 count less than a 100 you are at high risk of progression.

I do think what would be the best example of this is looking at the differences between America and the Europeans. America had an early epidemic and American patients are very heavily treated and therefore they are finding their options much more limited. They perhaps wanted their health to be perfect with drugs that weren’t particularly good. At the present time I still think we have this terrible balance between doing something now or doing something later. I think that is true also with salvage.

What to salvage from?

Sometimes you salvage from monotherapy but that is very uncommon. You may have dual NNRTI therapy and a few people are still on that. Three-drug HAART is the commonest but some people are only two class and remember it depends what you start on for what you salvage from.

If you use Trizivir, as an example, and it doesn’t work then you have already used up quite a few members of one class (nucleosides) although you’ve still got two classes left. It is a different situation to say if you failed on a protease and two NNRTIs.

You can salvage from toxicity, you may have someone who can’t take nucleosides, has already failed non-nucleosides and therefore is left with protease. There is quite a lot of shopping around but overall it is the multi drug exposure – it is the pyramid effect of the drugs you’ve taken and what experience and exposure you’ve had which will be lead to your salvage regimen.

Strategies for multiple treatment failure and/or multi drug resistance

Remember that these two things are slightly different.

You might have been infected recently with drug resistant HIV. If this is not picked up by doing resistance tests and careful monitoring of therapy you can burn through treatments very quickly and need salvage therapy very.

Do you continue treatment with a tolerable regimen?

There is very great uncertainty as to what the risk and benefit is. There’s been a couple of recent papers showing that over time resistance will definitely rise in people who have viral loads over 1000 copies/ml, but what is the consequence of that in terms of long term outcomes?

Should you continue or interrupt treatment?

Once again there is a lot of uncertainty about that which is why we are running the OPTIMA study.

There may be antiretroviral effects of a regimens residual activity. If you are failing for instance on ddI, d4T and nevirapine, and you’ve had PIs already, what should you do? Should you stop one drug? Should you stop them all? Should you immediately switch to another therapy?

Then finally there is the non-viral effects of a regimen and there are toxicities associated with the regimen. So you may feel more ill on something that might not be working for you than if you stopped. There are quality of life issues around that because you have to take the therapy and how it interferes with your life might be important.

Why there might be a residual treatment benefit?

This could be due to residual activity of treatment in spite of viral resistance i.e. if the resistance is only partial. That certainly happens with PIs and probably to a degree with nucleosides that you start of with low-level resistance that gets progressively worse if you continue on the same treatment.

It might be changes in viral characteristics due to that resistance i.e. that you have a less aggressive virus. There is a lot of accumulating evidence that some drugs may give you damaged or maimed virus – replication deficient virus – which would allow you to live in harmony with them relatively well even though the viral load was not completely suppressed.

There may be non anti-HIV mediated effects of treatment or even the effects on a different virus. You wouldn’t discontinue 3TC or tenofovir in somebody who has hepatitis B and HIV because they may be having an effect on their hepatitis as well. You would want to be very careful about those.

If you ask a virologist… John Erickson said three years ago, ‘The potential for generating more fit virus exists and requires no more than four or five successive mutations.’ So in a Calvinist way, ‘This should be a warning to those who advocate the widespread and continued use of PIs in patients with viral loads […] breaking through. A wiser course may be to cease protease inhibitor therapy in such cases to prevent the further accumulation of fitness mutations.’

Now that statement is clearly not coming from a patient or a clinician, it is coming from a scientist who is saying that we know these viruses will accumulate mutations therefore we must stop them. But of course what this screams out for is a trial to tell us whether that is the right thing to do because even though you might be developing protease resistant mutations, those mutations might be advantageous to you.

Ironically this was all before the replication capacity work came through suggesting that having a D30N mutation may actually not be that damaging in terms of the benefit you are getting from nelfinavir; or having an L90N mutation might still give you some benefit with saquinavir even though you do have resistance to it.

Nothing is black and white, nothing is necessarily written in stone, and everything I tell you today may be completely wrong in two years …

How effective are salvage therapies?

This is dependent really to a degree on the baseline viral load and CD4 count and the change in CD4 count and the options that you still have.

Competing factors are also important in terms of what you can gain so you have to have a balance between different things. Having a very low viral load and a very low CD4 may not be very much different from having a very high viral load and a very high CD4. Ironically you do see patients, especially some long term slow progressors who have a really low CD4 count but who are clinically very well because they also have a very low viral load as well.

Drug susceptibility and the amount of cross resistance you have is vitally important. However, as Clive has said to you, it is very difficult to evaluate how important that is and how you can evaluate it because the tests we have are not the best and are only improving slowly.

Previous drug exposure is linked to this and it very important. This makes me want to add ‘bad documentation’ the list at the start of my talk – I get many referral patients from other hospitals and inevitably there will be discrepancies between the drugs they say they have taken and the notes that are provided. When patients have moved from three or four hospitals you get to the position where nobody knows what they have had in the past, including them. That really is very destructive. One of things we are moving more towards is having an historical collection of information on every patient – but actually we need to have and it is something that could be a community initiative.

Patients could easily have a report of what they’ve had in the past which they keep themselves, there are ways of doing it, whether it is electronic or paper, it is going to be important in the future, especially knowing why people stopped therapy and whether there is an option to use that therapy again, what drug levels you can achieve and the number of classes available.

PK and pharmacology

PowerPoint Slide
click on slide for a larger view

The basic pharmacology in figure 2 (side 15) is important to understand for resistance and treatment failure. It shows why you aim to have drugs well above the level which will suppress 90% of virus.

The Cmax is the highest concentration you get after a dose. Cmin is the lowest concentration before the next dose. The area under the curve (AUC) is how much drug you get over the whole dosing period.

If you get into the yellow range (ie the levels of drug are below the IC90) before your next dose you get virus replication and this is what leads to treatment failure.

There many reasons why people get different response from treatment, some of which are very complicated. Some are genetic and we are more aware of how every one of us handles drugs very differently. We know that is true of alcohol. A Japanese person drinking whiskey will get drunk incredibly quickly because they don’t have alcohol dehydrogenase – the chemical which clears alcohol from the body.

Over time most of us have discovered that the more you drink the more you can drink because your body develops the ability to get rid of the alcohol. Alcohol is a drug, or is handled like a drug, and in this example it is no different to many of the drugs that we use.

The significance of this might be important depending on whether you’ve got resistance also. If you have virus that is not very resistant you may have quite a larger window of safety. However if you have a low drug levels because you are not absorbing it well – because of diarrhoea, or you may have missed the dose, or taken the wrong dose – then your safety margin goes down. If you have virus that is resistant a little bit then your safety margin also goes down.

PowerPoint Slide
click on slide for a larger view

This brings in the concept of the Inhibitory Quotient that is an individual calculation of the relation between drug levels and drug sensitivity in an individual patient.

If you get an increase in resistance and reduced absorption, or even normal absorption, you are at higher risk for resistance to develop. Over time those two things come together as you accumulate more resistance and as, perhaps, you absorb less drug, partly because your body is getting better at getting rid of it.

There is some suggestion that if you take a PI for four years, even if you take the same dose on the same day religiously you will have lower drug levels after four years than you had at the start because your body doesn’t like this. It is not normal to have nelfinavir inside of you so your body may develop new ways to get rid of it over time it. However it becomes very complicated because it is very difficult to measure some of these issues but they can be measured in cell culture, they can be measured in animals and in humans.

Figures 3 and 4 (slide 18 and 19) shown estimates of the Cmin/IC50 ratio from two pharmaceutical companies.

One drug company’s view of the balance between the lowest ever drug dose and the amount of drug required. They show that lopinavir/r is much better than indinavir, amprenavir and saquinavir… It is not surprising that Abbott produced lopinavir/r. If you look at Merck, we see indinavir/ritonavir is better and Merck produce indinavir. So the same, or slightly different, marker but it is the marker that is most useful for them, shows that indinavir and ritonavir works really well but lopinavir/r and saquinavir are rubbish.

PowerPoint Slide PowerPoint Slide
click on slide for a larger view

You have to be realistic about what you are seeing in terms of information.

PowerPoint Slide
click on slide for a larger view

I don’t know what the answer is but figure 5 (slide 21) is not a very good slide but it is real data. This uses patients isolates of virus with different types of mutations, so each one of these – and this one as you can see has a lot of mutations and these just single mutations. The indinavir level is up the vertical axis and shows the level of drug required to suppress virus completely. The top line shows the indinivir concentration boosted by ritonavir that you can achieve in the body. From this slide you would expect that every single virus, even multi drug resistant virus, can be suppressed by indinavir and it should always work. However, two not very big studies showed this isn’t the case. Moreno in Spain showed that after a period of time on treatment only 14% were undetectable and a third of them had kidney problems – so they got high levels of drug but it didn’t work. Brian Gazzard had the same with salvage patients using this combination with only 30% reducing viral load to less than 50 copies.

If the science tells us it is going to work it doesn’t mean it is going to work in practice. You have to question why doesn’t it work? Is it because the viruses behaves differently in those individuals? Is it because they couldn’t tolerate the combination? Is it because they had side effects? Or is it for other factors or a combination of those?

Many of these people who had not used a ritonavir boosted combination before. Other studies done with a 200mg dose or ritonavir have shown a relatively reasonable outcome in terms of suppression but certainly not a complete suppression.

Kaletra (lopinavir/r) is the most recent protease inhibitor. The 957 study showed 57 patients with extensive PI resistance (70% had baseline viral isolates with cross resistance to greater than three of the available PIs – so nasty resistance). People received lopinavir/r in combination with other drugs, but you still only lead to about 70% on-treatment effect. By intent-to-treat which is stricter and relates to what the patients were originally randomised showed about another 18% reduction – what you are saying is that 18% of those patients couldn’t take, or didn’t want to take that treatment.

If you had more resistance in this study you were less likely to have a response and if you had a lot of resistance, which was associated with already reduced susceptibility to new protease inhibitor then some people did not particularly well.

The phenotype really tells you what the capacity of that virus is also shows this reduction, baseline susceptibility, so if you have this fold change or increase of the level required to suppress then clearly you don’t get as much of a response over time. Then of course there is cross-resistance. Looking at the cross resistance between amprenavir and lopinavir/r, so if you take viral isolates with greater than eight mutations associated with lopinavir/r, if you look at that you can see there is a reduced susceptibility to lopinavir/r. There’s also a reduced effect on amprenavir – so you do get an effect on other viruses. However when you have rebound with patients failing amprenavir with changes – and this is looking at samples from patients – and this is getting reduced susceptibility which means when you have used one drug it inevitably has an effect on other drugs within the same class.

Other approaches

PI Boosting

We might be able to boost protease inhibitors, or use more than one, which we currently do for most people. Ironically, we will probably go back to using single PIs because I think that when people think of boosted protease they think they are using two but they are not they are using one – so indinavir/ritonavir is indinavir, saquinavir/ritonavir is saquinavir unless you are using greater than 800 mg per day of ritonavir. Because all you are doing is boosting the levels. Really we are still in the era of single proteases for most patients. You can use more nucleosides, an NNRTI or a nucleotide drug. You can use immune modulators and I think we are once again in on the very beginnings of using things like interleukin-2. Or you might have a treatment interruption and we’ll talk a little more about that. Or you can exhibit some restraint and try and prevent associated disease.

How should we protect people who are at risk?

A lot of the problem with bad doctors is that they have very little experience. Most junior staff at the Royal Free have only seen one or two people with cryptococomeningitis, some of them have never seen toxoplasmosis, most of them have never seen CMV retinitis. I used to diagnose one of these OIs every three or four days and would follow them over time. A current research project with UCL on Kaposi Sarcoma (KS) in unable to find people with lesions so they can biopsy. Almost nobody has KS and it is a major issue when a patient presents with something unusual, that a doctor has not got experience with.

With prophylaxis for opportunistic infections (prophylaxis are drugs that you take to prevent an infection) there are a huge number of studies that tell us what to do. Prophylaxis should be started at specific CD4 thresholds. Almost all of the opportunistic infections that have occurred in structured treatment interruption (STI) studies should have been preventable. It has been bad management that has resulted in the difference in the study rather than actual changes.

  • CD4 less than 200: you should think about PCP and toxoplasmosis prophylactics
  • CD4 less than 100: MAI prophylaxis. I have just seen another case of someone who was given treatment with a low CD4 count, wasn’t MAC or MAI prophylaxis and ended up with MAI in their neck, totally unnecessary, you can prevent that just by putting in some antibiotics for twelve weeks.
  • CD4 less than 50: you need monitoring of CMV viral load, which is now available, and treatment if positive, regular review and good contact. It is all about good management external to the treatment.

Dual or triple boosted PIs?

Dual or triple boosted PIs are probably they are more effective and more durable than single drugs. They usually exploit the change in PK and they can result in reduced pill counts. I think one of the more exciting things that are coming along is that nelfinavir and saquinavir will actually be many fewer drugs than they were. Atazanavir is only two pills once a day compared to currently saquinavir five pills twice a day. There are constant PK improvements but it is up to us to pressurise the companies to continue to do that development. They are very slow at doing that compared to working on a new drug because it doesn’t often improve their market.

The cons against that is that you do have potential cross resistance so just improving the number of drugs that you use or expanding it may not always be the best thing to do. You do get additive toxicity so the more drugs you add the more likely you are to get side effects.

You may enhance the nucleoside regimens – Trizivir improves the delivery – but also you are increasing the AZT dose so you might get more toxicity. There is the use of combinations some of which have been seen to be beneficial, some of which like ddI, d4T, hydroxyurea might slightly improve the surrogate marker outcome but definitely increase the toxicity. There was an example of where the toxicity downside was worse than the upside.

Abacavir is a relatively new nucleoside and tenofovir was approved this year, and there will be more.

The development of a drug called aluvudine was in the early 1990s because it had a lot of side effects at the dose used. It has just been used in a salvage study by the French and appeared to be, in the short term, quite successful. It has a very different resistance profile. What may be beneficial for some people is to use that drug for a short period of time or at a low dose.

Mega-HAART (see slides 38-53)

What is the rationale of Mega-HAART for HAART failure?

Q: Can you define mega-HAART?

A: Mega-Haart is not exactly defined. It is greater than five drugs probably. Basically mega-HAART is ‘throwing the kitchen sink in’. With higher numbers of drugs (some Mega-HAART regimens use up to nine drugs), you can use 2-3 PIs, 3-4 nucleosides, an NNRTI and hydroxyurea. It is very ill defined because it depends I suppose on what people’s perceptions of their pills are. For some people mega-HAART would be more than one pill twice a day. For some of my gym bunnies the more pills I give them the more they like it. I say to them. ‘We could reduce the number of pills’ and they say, ‘No, no I like the fact that I am taking lots of pills’, and they’ll take thousands of vitamins. I do think with peoples health you have to work out what they do like and what they don’t like. Do you like eating with your pills? do you not like eating? do you like taking things early in the morning or late at night or do you want to do everything at lunch? Those ar ethings you have to look at.

We ran a mega-HAART study when efavirenz was first available and we had many patients at that time who had been relatively experienced and had very few new drug options other than efavirenz. Although nevirapine had been available there was a perception in the early years that it wasn’t a particularly good drug based, really, around one very bad study that was done by Boehringer (1090 study). They took failing patients at late stage disease, added their drug, and it didn’t really show any benefit. Ironically nevirapine is quite a good drug and in certain situations probably better than efavirenz but in that setting it was not.

For most people efavirenz was a therefore a new class of compound.

We took 92 patients, most of whom were NNRTI-naïve (only 16% had used nevirapine), and we gave them a salvage regimen. They were mainly gay men, average age 37. Most had a low CD4, 26% had less than 50 CD4 cells and 75% were less than 200. Most people had very high viral loads – median 310,000 viral load and only 10% under 10,000. This was an advanced population and most people had had AIDS. We had a lot of sick patients then which is not the case now.

They had used a lot of drugs, but only one or two protease inhibitors because those drugs were not so good then. Most people had used 3-4 nucleosides. They started on efavirenz plus a whole raft of different things. Importantly we individualised combinations based on what they wanted to do (we didn’t have resistance testing) but also designed something that worked. I would say, ‘Well if you think you don’t want to take four or five pills then you probably ought not to do this and stick as you are rather than wasting an opportunity.’

What I found most useful in this setting was getting patients who had taken therapy to talk to other patients. We don’t utilise that peer support system as much as we should. We have done it recently IL-2 and now patients on T-20 are now being used as advocates to talk about how they cope with the injections.


  • The vast majority of patients got a good response, which improved over time.
  • Over one year only a small number of patients died.
  • Over 50% got viral load below 50 copies.
  • Many people were fully adherent for the first time in their life, they had taken lots of regimens, but never really taken it seriously.
  • The strength of the combination was important but it was also due to the strength of the resolve to get a benefit from treatment.

What happened in the second year for these 92 patients? Did they get worse? How durable was this effect?

  • More people did die, several from tumours, one from hepatitis C and I think there was one suicide and one accident death, whether those were related to HIV is another matter. But people do die, and this rate is probably still low for a very advanced group.
  • But, there was no real loss of anti-HIV effect.
  • About 50% still have less than 50 copies. Undoubtedly in that group people have changed therapy, we’ve plotted their change of therapy, but they have maintained what the outcome was that they achieved in the first place. Certainly there was a difference if you’d already had resistance so the efavirenz was clearly an important issue in that combination. If you had previous nevirapine you didn’t do quite as well, but this was relatively small numbers of patients.
  • CD4 continued to rise and continues to rise in those patients. If you look at the difference between year one changes in CD4 – 36% had greater than 100 CD4 rise in year one and over two years, 50% greater than 100 CD4 cells.
  • It was actually an improving salvage regimen rather than a deteriorating salvage regimen. Those sections of patients with lower CD4 responses or – that is one where they’ve got a negative CD4 response – hasn’t really changed over that time period. There was some effect of hydroxyurea on blunting that because it does have an effect on blunting CD4 rise that was just significant.

Why did that work and why did it continue to work?

  • One new class and agent was added to most patients so they all got something new and this explains a lot of the benefit. In today’s terms that might be T20, or it might be another receptor blocker, something that they haven’t had before.
  • There was restrictive prescribing – there were only four doctors doing this. At the Royal Free only senior staff can prescribe antiretrovirals. One of the big problems is when physician change a drug that they have no real evidence, or no real experience, with.
  • We closely monitored and cared for these patients – because they are the group of patients who are most at risk. Most of my patients see me for less than 5 minutes, but half an hour if they actually need it. Most of the time I need to spend more time with people.
  • The dual PI had an effect, or at least the boosted PI.
  • Treatment interruption? patients who had treatment interruption before they started on therapy did slightly better but that has been diluted out by the treatment effect now

Julio Montaner has a very experienced cohort in British Colombia that got similar results although their intention-to-treat results were was not so good (slide 53). This was partly for two reasons. One is that they have a much more chaotic group of patients, they have a lot of active drug users and that really has been seen in a lot of settings to mean that people are less able to access care. They also have viruses that are probably a lot more resistant in their population.

Q: Did you have people on 800mg/400mg combinations? (indinavir/ritonavir)

A: Yes, we started them off on 800/400mg – pretty hideous, very hard work. Even 800/200 can be difficult. We gave them Ondansetron to control nausea and some people were treated as in-patients in hospital. This was like the times in 1995 when I had people who were getting indinavir for the first time and I rang everybody up and said come into the clinic tomorrow. It is the same effect. For somebody who has got a low CD4 count you can’t say ‘You’ll be fine with 30 T cells’ – no, it is a medical emergency to have fewer than 50 T cells even if you haven’t got an opportunistic infection.

Q: Clive showed a very early AZT study earlier, high dose 1500 or 2000 showing a three to four log reduction. Do you use higher doses to get greater anti-HIV effect?

A: Yes, and it may be possible to give high dose tenofovir. The next salvage patient is going to be on TMC125, saquinavir, lopinavir and high dose tenofovir and possibly delavirdine because he previously had lactic acidosis and we are very cautious about nucleosides. I think there are some things initially you don’t think are worth doing. We intensified a lot of these patients by adding 3TC and we kind of went up and down with some of the regimens so they’d stop some things for a few weeks and then restart them. There is a lot of changing and constant contact with people. We’d chase people because otherwise some people get lost and just don’t necessarily get the treatment they need.

New strategies and drugs

This is not an exhaustive list but limitations at the moment include:

  • resistance and cross resistance,
  • the complexity of dosing – although arguably now most drugs are going to once or twice a day,
  • tolerability
  • drug interactions – especially with things like hep C patients
  • there is limited data on organ tissue penetration but quite a bit of data suggesting there are reservoirs that drugs don’t get into easily,
  • unmeasurable or verifiable PK – we are only just getting a test to look at the effect of ddI and tenofovir in cells and tenofovir
  • cost.

What is the ideal regimen? Well we want something that completely eradicates HIV from the reservoirs and suppresses it completely; you want no side effects, no resistance. You want it simple – one pill once a day, or once a week, or once a year with no food restrictions; and you want something to boost your immune system. That is what you really want.

Novel compounds in the pipeline

Fusion inhibitors are opening a doorway, how wide that doorway will be is another matter because these drugs have to be given by injection. Roche need some recognition for having developed T-20 because it may never make them a profit. It seems to have done something to their share price though which I don’t really understand. It may never be widely used though it may open doors to other areas. They have pushed forwards with it quite interestingly against all of the economic prerogatives of companies. Slide 60 shows how T-20 works, but there are several nice videos that show much better how T-20 works on the internet at the Roche site: (

T-20 is also a very complex molecule – slide 63 show the complexity compared to AZT. Saquinavir that is much more complicated than AZT and takes a long while to produce, but T-20 is a nightmare of a compound to produce. Roche have at least been honest and said this. The question is the balance between how fast you can produce it and how many people need it. In this country we still have T-20 slots open in our studies – although this may be because the people who really need it are not being directed to these studies – but we still have slots open and it is not going to be a major issue about access.

Chemokine receptor antagonists might be more acceptable and there are some very exciting compounds here. Shering has one that they are going to bring to the UK in phase II studies by the beginning of 2003. Pfizer-Agouron approached are going to be doing a phase 1 monotherapy study, hopefully starting before the end of 2002. That is controversial but I think the data suggests that resistance is unlikely over such a short period of time. Data from the Shering compound suggests that you don’t get resistance so I think the Pfizer one may be similar.

Zinc finger inhibitors haven’t really gone anywhere at the moment.

Integrase inhibitors – where you are inhibiting the ability of the virus to insert itself into the nucleus or the brain of the cell, were quite exciting a few years ago but the first compounds that people looked at were actually not integrase inhibitors in their totality and they were doing some other effects in the cells. Newer compounds which were presented by the Merck group and Shanogi seemed to be more effective and may be in clinical trials in the middle of 2003.

Q: What is zinc finger?

A: It is partly to do with the way the virus gets made and it is inhibiting the packaging of the virus. There are some drugs in test tubes but no drugs that have reached anywhere that it is going to be particularly useful.

Immune modulators? IL-2 continues to be interesting but it is not easy to take and I think that is the biggest problem with it. Also how much benefit do you get over and above what you get from drugs? Clearly it doesn’t work if the drugs don’t work, so it is not a salvage drug. Gm-CSF is more interesting and there’s survival benefit. Gm-CSF is something that stimulates your white cells which are not directed against HIV so they are directed against bacterial infections and things like TB. It is almost like making the best of what you’ve got left but then stimulating the production of that, it is like fertilising your roses. Because it is given by needle and because it is expensive people don’t use it when actually they should. But there are certain situations when it is beneficial and there is a survival study in people with HIV and a CD$ count under 50 that showed benefit. It is also used in people with chronic TB.

Protease inhibitors – tipranavir is a drug that has potential but a dose has still not been clearly defined. Boehringer are having some difficulty with other drugs and therefore they may not be seeing this as a huge priority. It is a pity because the results looked promising for people who had failed one PI – although they weren’t wonderfully better than saquinavir. However tipranavir does have an advantage in that it seems to have no cross-resistance with the available protease inhibitors.

Q: T-20 is a drug that works against the virus?

A: Yes, it works against the virus – and is the first HIV drug to do this – so T-20 doesn’t really have any side effects, other than those due to the process of giving the injections themselves.

Q: What does the MW stand for?

A: Molecular weight. So your molecular weight here is 4400 against 200.

Q: How would someone access T-20 now if they are not treated at one of the trial sites?

A: People can access T20 currently, and I know there is one trial place still open but there will then be small numbers of expanded access open in larger centres. Anybody anywhere other than those sites needs to contact somebody at those sites. This is essentially that is true of all trial and expanded access drugs. If you want a new drug you have to get it from a big site and that is partly because that is where all the work goes on. It is like any serious medical condition – if you have malignant melanoma you don’t get looked after in your local district general because there is no experience there. That is the difficulty for people who live outside of London, at least in London it is only a tube ride away to go and see somebody to access. If you live in Dorset it is much more difficult. That is why, on the whole, people outside London have more problems with their therapy because of lack of access. It is a difficult issue.

Q: Is there resistance to T-20?

A: Yes. The initial safety feasibility study was an add-on study. Slide 65 shows that if you were resistant to other drugs it appeared there was no cross resistance to T-20 or T-1249 (the second-generation fusion inhibitor). One of the difficulties was that there was no assay for fusion inhibitor resistance but now Virologic have developed one. Once again you can’t test resistance if you don’t have a measure to test resistance and that is now available. There was some data published in the report I did on HIV and Hepatitis ( that looked at T-20 and T-1249 resistance. Basically people who developed resistance to T-20 were all sensitive to the se T-1249. The evidence from studies strongly argues that if you don’t need to use T-20 today then don’t use it. You should wait until you have at least one partner drug i.e. for people who have got a lot of resistance T-20 is probably not the right drug to take without another drug – and that is why some people have failed. The studies show that if you only add in T-20 you don’t do wonderfully well.

Results from two big T-20 studies were presented at Barcelona (slide 66-69). They were all patients failing on at least two PIs and one NNRTI and one RTI so it is three class resistance. They had viral loads greater than 5000 so these were real failures as opposed to just blips. It is big numbers so you’ve got 300 people on T-20. Optimised background meant whatever you could take or whatever you and your doctor thought you were sensitive to still and it wasn’t much defined beyond that. Lowish CD4s, under 100, although not necessarily really low. What is interesting is that with every study, as time goes by, we get greyer, shorter, older and so does the average age of the patient. Most big studies used to be a median age of 32-33 and now it is 42-43. Viral load was high at over 100,000 copies/ml and they were very treatment experienced.

This means that if you have 100,000 copies and you can’t add to this combination you are going to go down to 10,000 copies and get resistance and go straight back up again. Probably that is the wrong scenario to use it in. If you looked at responses by week 24,greater than one log decrease you are getting about twice as many people, less than 400 copies the same, less than 50 copies about 25%, There is no difference between TORO1 and TORO 2 they just happen to be one in Europe and one in America.

Q: Is T20 more effective with certain other drugs?

A: Nobody knows. There is no subgroup analysis on this study at the moment or at least I’ve not seen one, which looks at if you get more benefit if you happen to have a new PI versus a new NNRTI. You wouldn’t have thought that would have an effect because there is no scientific interaction between the different parts of the virus cycle.

Q: Will T-20 mean a lower burden of other drugs, can you take fewer drugs with it?

A: It depends what drugs you are going to take, but in salvage therapy you are need a strong regimen to support it.

Q: I’m talking about it as an add in.

A: You could substitute it but we are talking about a salvage scenario.

Q: After the salvage situation.

A: After the salvage situation – undoubtedly and one idea is to take T20 as monotherapy at seroconversion and we’ve not gone beyond that idea. T20 may be a very interesting drug to take as you are seroconverting.

Q: Could you use it as PEP (post-exposure prophylaxis)?

A: Although we have a quite a big concern that PEP doesn’t work, a fusion inhibitor might work better because if is actually blocking entry itself. Co-receptor blockers may work better because they are absorbed but it might be worth having a shot of T-20 sitting in your fridge in case the condom breaks. Actually that is not the way most people are dealing with it as most people are not using condoms so they can’t tell when they are at risk. If we move onto the safer sex subject, that is the difficulty with unprotected sex, at least if you have some protection you know when that protection is not working for the most part. If you have no protection then the factors that are involved in whether you seroconvert are outside of your experience i.e. they are involved in the viral load of the other person or what kind of virus they’ve got rather than partly your protection.

Q: Is you comment that most people aren’t practicing safer sex based on your patient group?

A: No it is based on a recent gym survey – sorry I should have said ‘most gay men’.

Q: Very first studies in T-20 also showed a dose effect so where you didn’t have additional drugs and you have access to more T-20 would you increase the dose?

A: No.

Q: It T-20 is going to be a very expensive drug what do people do if they are in a health authority that doesn’t prescribe it?

A: Go to a health authority where you can get it i.e. come to London, that is the bottom line, or sit on the commissioners desk or go and see Tony Blair, or be active. We are having that discussion at the moment because there are acute seroconversions of hepatitis C. We are having to make individual applications to the health authority for funding on a patient-by-patient basis. None have been turned down yet but that will cost probably a similar amount to T-20.

Q: How often is T-20 given?

A: Twice a day.

Q: So you need two injections every day?

A: It depends; it can be once a day, it is going to be once a day eventually for everyone. In different studies it is different because some of them are doing higher doses.

Structured treatment interruptions – why would you do it?

Cost, side effects and perhaps immune response. There is some evidence that you get better immune responses with people on HAART versus long term non-progressors, so re-exposure to HIV during intermittent treatment might stimulate your immune response i.e. when you are on treatment you kind of forget that you are trying to fight.

Types of structured treatment interruptions:

  • During primary treatment, – this is very exciting, but probably not on its own in terms of effect and we now looking at giving vaccines before we stop therapy.
  • With people on effective HAART starting during chronic infection
  • With people who are failing.

Reducing viral load doesn’t seem to happen – if you stop and start, stop and start therapy you don’t get a declining viral load over time and the and there’s not evidence that that is a particularly good intervention. In terms of what your benefit might be in terms of a chronic setting with people who are failing therapy it may be that you are allowing the virus to reverse to a more wild type virus.

This is what we are talking about with replication capacity. This is changing drug susceptibility during a treatment interruption. Over time after you’ve stopped therapy you actually get this reduction in drug susceptibility so you might get a better benefit when you start on therapy afterwards and this can be seen in some of the Steve Deeks studies. After a treatment interruption your viral load goes up but your CD4 count can remain stable for a while before it then goes down. When you switch your virus becomes more aggressive again. Drug resistant virus is a ‘softer’ type of virus – therefore it doesn’t have that nasty effect on the viral load, the CD4 count goes down, but it doesn’t go down as rapidly as it does when the virus becomes nasty again. It is the blunting effect of the virus.

The OPTIMA salvage study is particularly unsuccessful at the moment because there aren’t enough patients. The idea is to try and work out strategy with people who have only limited options. When you eventually don’t have any more options you have to decide what to do with the available things:

  • should you give people more drugs or less drugs when you are switching therapy and
  • should you give them a break or no break between switching therapies.

It is aimed at people who have already failed two drug regimens and it is trying to evaluate mega-HAART versus standard HAART and whether having a drug free three-month period is better than immediately switching. It is being run in 75 sites in Canada, UK and USA.

People get screened and randomised into this drug free period or no drug free period, so an immediate switch to new treatment or having a break.

Then into these different groups and the doctors and the patients can then decide between them what they do. They get a phenotypic drug resistance test to work out what their best options are. It is aimed at people with lowish CD4s although this is the older criteria and they have now been improved so you can now have people with CD4 counts up to 300 but with a high viral load.

It is people who have documented failure as opposed to just a blip. The problem we have in the UK is that we have very few patients, at least in the centres that are doing studies. I get people referred from outside of London from smaller centres and really they are in a real mess but nobody, including the patient, felt it was appropriate to get them somewhere where things might be different. Why I think this is important that we’ve seen this effect in other studies before. Caesar was an old study that looked at continuing failing treatment treatment but adding in 3TC – or adding in an early NNRTI that we don’t use any more.

This is why it is important to do studies in people who are not completely suppressed. You got a 50% reduction in death rate and AIDS in people who added in a treatment to which everybody became resistant to immediately, or within 12 weeks. So even if you are adding something in and it doesn’t seem to be working it may give you a beneficial outcome if you follow up people long enough. I think with some of the studies we’ve seen, even the 20 studies in patients who are resistant to 20, they may still benefit over time and some of the shorter studies may not easily define that.


  • The conclusion is that salvage therapy can and does work.
  • Determinants of success are becoming clearer and the patient is the major factor.
  • It is unclear whether to wait for new agents or classes and it is very important to have a multi-disciplinary approach.
  • Some recent infections may need salvage soon and I think this is one of the worrying things that people who are recently infected may be progressing faster.
  • Prophylaxis is part of patient management and timing is everything – and the way you use them is important

I’ll finish with the observation is that if you can stay calm while everything around you is in chaos you probably haven’t understood how serious things are….

Published: August 2, 2002
Last edited: January 17, 2011