NIH CLINICAL CENTER GRAND ROUNDS
Episode 2009-015
Time: 57:13
Recorded April 29, 2009
Genetic Variation of HIV-1 Populations in vivo: Implications for
the Emergence of Antiretroviral Drug Resistance
Frank Maldarelli, MD, PhD, Staff Clinician, HIV Drug Resistance Program, NCI
Recent Findings in the Epidemiology of Cancer in HIV-Infected People
Eric A. Engels, MD, MPH, Senior Investigator, Infections and Immunoepidemiology Branch
Division of Cancer Epidemiology and Genetics, NCI
ANNOUNCER: Discussing Outstanding Science of the Past, Present and Future – this is NIH Clinical Center Grand Rounds.
(Music establishes, goes under VO)
ANNOUNCER: Greetings and welcome to NIH Clinical Center Grand Rounds. We have two speakers on today's presentation, recorded April 29, 2009. Dr. Frank Maldarelli, from the HIV Drug Resistance Program at the National Cancer Institute will discuss Genetic Variation of HIV-1 Populations in vivo: Implications for the Emergence of Antiretroviral Drug Resistance. He’ll be followed by Dr. Eric Engels, Senior Investigator in the Infections and Immunoepidemiology Branch of the Division of Cancer Epidemiology and Genetics, NCI, who will present Recent Findings in the Epidemiology of Cancer in HIV-infected People.
We take you to the Lipsett Ampitheater at the NIH Clinical Center in Bethesda, Maryland, where Dr. John I. Gallin, director of the Clinical Center, will introduce today's first speaker.
GALLIN: Okay, good afternoon. Welcome to Clinical Center Grand Rounds. Today we have two speakers who will speak on the topic of HIV.
Our first speaker is Frank Maldarelli, and his talk with talk about genetic variation of HIV-1 populations in vivo. Let me tell you about Frank and then he’ll speak and then I will introduce our second speaker.
Dr. Maldarelli is a staff clinician at the HIV drug resistance program at the National Cancer Institute. Since 1998 he has headed the in vivo biology group. He received a baccalaureate degree from Johns Hopkins and then received his Ph.D. from City University of New York and his M.D. at Mt. Sinai. After completing his residency in internal medicine at Columbia, he joined the laboratory of molecular microbiology at the National Institutes of Virology and Disease as a medical staff fellow and since 2003 he has been PI on five bench to bedside awards. He's a member of the Infectious Disease Society of America and his current research interests include HIV-one population genetics, antiretroviral drug resistance in mechanisms of HIV-one persistence and it's my pleasure to welcome him to the podium.
[ applause ]
MALDARELLI: Good afternoon, I think anybody who looked at the yellow sheet for the last couple of weeks knows there's quite an interest in HIV and at the retroviruses at the National Institutes of Health. And it's a relatively tight community and one that has lost a particularly valued member in the last couple of days.
I’m not sure ... And that's John Brady. I don't know how many of you have known him, but a great man, great mentor and a great scientist. Been here for more than 20 years and I think amassed quite a pile of information on retroviruses. So, if you hadn't seen the message already from Bob Wiltrot, you certainly will and we will miss him.
The story that I wanted to tell him and the rest of you about today has certain objectives, understanding HIV, population genetics and how understanding that may help us to better understanding of the development of drug resistance in individuals. I think I have no relevant financial relationships or conflicts of interest to talk about.
Drug resistance is an important concern for both from an individual and population basis. For an individual, resistance results in rebound viremia and death if it's there, and it limits therapy and becomes much more resistent. On a global level, multiple regimens are more expensive for treatment and resistance may spread at the time of transmission. So those concerns are estimated here for individuals about 20-40% of patients who will develop drug resistance on newer ones with we hope it will issue much less and on a global level, transmitted drug resistance about 15% for the NNRTI class and 5% for (…) and 5% for PIS and this tends to refer to all the gains realized by the drug development of the last 20 years.
And here's one illustration of that, this showing the global burden if you will, in excess of 33 million cases in the world and it's relation to the roll out of antiretroviral therapy so certain countries, specially those that are experiencing an increase in their epidemic spread of HIV are also attempting to spread drug therapy to those population and in those places like Cambodia, China where the retroviral coverage is going to increase to 30 or 5, to about 25%, the spread of resistance may be even more considerable than places where the epidemic is relatively well established, places like South Africa where there are many infections already to start with.
So spreading drugs and spreading resistance may go, unfortunately hand in hand. So understanding that in some detail, there are many causes for drug resistance. Obviously drug adherence is going to be a major one but understanding that on a basic level is one of the things that the HIV drug resistance program does and I think the point I would like to start with, is that HIV replication on a basic level is both rapid and error prone and perhaps you've heard those bandied about before but the error prone nature of antiviral retrovirus system something like one mistake every time it replicates its genome. Coupled with a rapid replication rate, something like once every one to 2 days, results in a very large and diverse population of viruses within an individual from which drug resistance can emerge and this, this entire cycle here probably takes place, reverse transcription in seminal work done by Malcom Martin's group can occur in the first few hours of infection and then new virion production within one to 2 days even within infected individuals.
So what are the consequences of that for the development of drug resistance? And I guess the thought I’d like to disabuse you all of is that this rapid and error prone mechanism is somehow a bunch of mistakes that a virus makes because it's some sort of primordial mechanism and it doesn't know what it's doing and I continuing knows very well if it can think what it's doing and this rapid and error prone mechanism is actually a pathogenic determinant and here's one example of that.
So if each round of HIV generates numerous mutants, the ability of those mutants to replicate is going to vary greatly in terms of their fitness. Some will still be tolerated at a low frequency. And I diverse population is a single amino acid within the reverse transcript ace of HIV 184 is normally methynine, a single nucleotide can make that into a GTG, coded valene or a ATA encoding isoloosely. In this form this reverse transcriptase is sensative to all retrovirals. This one and this one completely resistant to several antiretrovirals that are used in the clinic today.
So the fact that these may exist at a relatively low level, they're tolerated, they replicate but they can be selected quickly and dynamically in the presence of drug selection pressure. So the presence of one of those mutations in RT is relatively likely, 2 less so, 3, somewhat statistically unfavorable but individual mutations may combine fairly rapidly through a combination process that occurs during replication.
So I guess the point, maybe the most important point for my 20 minutes is that drug therapy is a selective pressure that permits resistant viruses that preexist to emerge. Nice thoughts, originally described by people here at NIH and John Coffin in another science paper in the mid-90s but is that actually detectable. So using a technique described by Sarah Palmer, called PCR, we can actually detect low level mutations. Low level being somewhere around 3 in a thousand. And in a series of patients seen here in NIAID we example low level mutations confirming resistance to the drug of fabrics, position k103 n. And most patients had never seen or been exposed to this antiretroviral, the relative level of that mutation was at our level of detection. But in a couple of of individuals. These 2 and in this patient, twice, that mutation was present even though the person had never been exposed to that antiretroviral. Providing the first proof, really that preexisting mutations can preexist prior to therapy.
So it's no surprise then that preexisting mutations can have a strong impact on therapy and those of you who know the speil are very well acquainted with these data and that's that single dose of navarapine to prevent mother to child transferance, so a single dose of nevarapine using that same all specific PCR test, we can find approximately 83% of all the women who receive that in certain trials had drug resistance to navarapine detectable within one group after 2 months or after 4 or after 6 months. Once again this was re-enforced in the octane trial but many of you know those individuals in Africa who had received single dose navarapine who subsequently required combination therapy to control HIV infection were randomized to get a pi containing regimen or an NNTRI containing regimen and those group, those individuals who had gotten their single dose navarapine to prevent mother to child transmission between 6 to 12 months had failure at a rate of 37%, 12-24 months after getting that single dose of it, at 24% and even after 2 years after getting that single dose of of navarapine there was still some evidence of navarapine resistance, causing this trial as many of you know to be stopped by the DSMB.
So understanding how that drug resistance may occur is a function of understanding how reverse transcript ace makes errors and how fast HIV replicates but understanding how that mutation may spread in an individual is really a population genetics question and one that we've been very interested in. So, one of the things we've been trying to do is understand that that development of resistance in terms of the entire population. And to do that, we've enrolled a number of individuals here, once again in collaboration with NIAID in several clinical protocols and in this case all HIV infected adults and antiretrovirals outfitted from days to years, and any retroviral therapy had points as prescribed by current DHHS recommendations.
For these individuals we sampled them on a daily and then weekly and then out to 18 months and then in a prolonged period out to several years. And so what I would like to describe then is what happens to these HIV populations upon infection and then over time prior to therapy and then what happens to them on therapy.
So this is one of the techniques that we used to understand HIV populations and may many of you may be exposed to it before, philogenetic analysis, so looking at the differences in nucleotides in pair wise comparisons. And those of you who are familiar with something like that, know that we normally compare them to a standard laboratory strain, we also use pnl 4 3 and we know (…) compared them at the NIH. And this individual who was infected 20 days before the first sample occurred, the number of individual verions that we sequence somewhere around 20 were genetically nearly all identical. So this is using a technique known as single genome sequencing, which we can identify the sequence of individual virions and in this experiment, early in infection, we found a very monomorphic virus population and those who are interested may see a recent publication by George Shaw that described this in much greater detail, showing the transmission of a single HIV virion in the majority of cases of HIV infection.
So things start out fairly boring and then with time, these founder effects can determine the genetic composition of the initial HIV composition and here's a series of 11 and then an additional 12 patients who were infected for varying times from whom we obtain samples and determine their genetic diversity or average pair wise difference. And here they are described as their time after infection. So, the time from the point they were infected early in infection as you see, the average pair wise difference for genetic diversity is quite low and then with time, it increases, and then, it gets over 2 and half percent in the region we're looking at which is the polymerase gene and encoding protease and reverse transcript ace. And so that early period actually approximates the accumulation of change in this early period approximate the RT rate so it's recapitulating molecular biology, but then this additional accumulation of diversity seems to be more of a balance between the ability of the virus to generate and fix mutations, and selection of the host to be able to suppress it.
At an end time we can find a fairly complex philogenetic tree, much more complicated than the one that we see in the early period. And here's that on an individual basis and I think the thing I wanted to point out was that the emergence of genetic diversity is low but perceptible, so here in one individual seen here at the clinic, over several years, HIV RNA levels remain relatively stable and if we sample them at the times indicated by color, we can see that the diversity in that individual can increase from a relatively low answer .4% to above, about one.3%. And in second example of that same thing, over a longer period in somebody with a more established infection, this increase is quite slow but still detectable and the philogenetic tree that we see here is you know it's sort of nice to look at. It's got a lot of nice colors in it. It makes you think that boy all these things are separated on different branches and they all have plenty of individual differences but the thing I would like to get the other point I would like to make is that this is just almost a fact of the appearance. How many of these sequences are truly and significantly different, one from another, using a boot trap approach, the only ones that are really different are here in the thick bars, so these, yeah, they're different from those, these, they're different from the rest and these and that one are different from the rest and I think the thing I wanted to get across is that overtime genetic composition of that HIV in an individual doesn't change all that quickly. And this technique of philogenetics really can't display that all that well.
And so, one of the things that we tried to do, is to develop methods to detect change, is it fast or slow. And you know from the virus standpoint, the virus replicates somewhere around 1-2 days. So generations are a lot faster than ours and consider how things don't change in our own population. So here, are 2 individuals who were painted on almost 500 years ago and here a summer student, right here at the NIH and some things really don't change over a period. They all say the same thing: "I am smiling." So we needed new techniques to try and distinguish that rate of change and the one we developed with Harvard was something called a nonparametric test for the significance of genetic distances and this is based on a test by Hudson for geographic subdivision and basically we try and decide whether 2 samples taken at 2 different time points, so one group and a second group, are they part of the same population? Or are they part of different populations? And what we do measure genetic differences between the 2 genetic populations and then scramble them and re-measure that distance. And do that 10,000 times. And come up with a probability that these 2 populations were drawn the same or different population? In other words the more negative the number of the things we're going to look at, the more the chances that this thing, these 2 populations were drawn from a different set of from the different population.
And so here's that individual that we saw before, you know, fairly stable, viral RNA level, is fairly stable, diversity but when we look by this geographic subdivision test, we see that over time the probability that these samples were drawn from the same population goes way, way, down and after about 3 or 4 years, the chances that they were drawn from the same population is less than 10 to the ninth. So although those populations individually are genetically stable, they change relatively slowly. Something like every 1100 days and that relative to the generation time of the HIV virus is actually quite simple. So it's monomorphic on its section, bi phasic in diversity on accumulation of change, with a change that sort of levels out somewhere around 2% and these populations are relatively large. What happens on therapies.
So those who follow this field are familiar with a curve like this, with time viral RNA level on therapy declined to a stable low level with several phase declines. And here an individual seen at the NIH, we find a similar decrease in viremia on starting antiretroviral therapy but strikingly the diversity does not change. So if you look at diversity both before and after therapy, you will see a fairly stable level of genetic diversity, somewhere starting at one.4 and ending at one.4 and if we look at the philogenetic distribution of those sequences, we find that they continue to be fairly well inter-mixed so whether you are at a viral RNA level of close to a million or 10,000 or a hundred thousand fold lower than that, the genetic diversity remains relatively stable.
Using our pan mixia index, we also find that there is no change in the genetic composition by that measure, in most patients. Next. In one final individual that I just would like to bring up quickly, we can see after a long period on therapy, the emergence of what others have described as a dominant plasma clone. In other words an individual sequence that's repeated more than a few times. See, evidence of that here and here? Suggesting that there is slow change in the virus population over time but in general, and here's a box of plot, containing pre-therapy to on therapy of no patients there was no significant change in genetic diversity.
So our model suggests that there's a similar philogenetic structure, a common source of HIV genetic diversity present in those cells that are infected that turnover quickly, or that turnover slowly. Therapy doesn't initially select for restricted group of HIV, there's no genetic bottle neck initially and the persistence of diversity implies that preexisting mutations including drug resistance mutations are unlikely to be eliminated by antiretroviral therapy.
So I’ll just quickly pass through this and then just acknowledge the individuals who actually do all the work. Much of this was done by Mary Kerny and Sarah Palmer now at other locations, and with our program, here, Kristin who worked on the genetic analysis, Francesca and others who are working working to recruit more patients and then this is a wonderful collaboration we had with NIAID studying patients with Mike Bowls and Joanne and Joe Kovax, in Rick Davey’s program and with Cliff Lane on the eighth floor. John Coffin started this set of experiments and dreamed up the analysis and directed this work while me was in charge of the drug resistance program now followed by Steve Hughes and we have marvelous collabrative help and I’ll stop there because I think I’m a little over my time. Thank you.
[ applause ]
GALLIN: Thank you. We have time for some questions. Henry?
[ inaudible question ]
MALDARELLI: So, Henry's trying to tell me I should have been talking about this for the last 20 minutes and you're probably right, but, so, the fact that we can describe genetic diversity in an individual, we can put all of those individuals into a similar kind of analysis, and see whether or not they're essentially related and this has been done in other geographic locations and notably Europe and the United Kingdom so it's actually possible to track an epidemic in terms of the transmission network based on genetic related viruses. It's not a well described technique yet but it's something we're interested in doing obviously in collaboration with Henry and so, we've started an analysis just like this, studying patients who are seen in the NIAID CCMD clinic upstairs and what's known as the fellows clinic so patient who is are seen here in the area, probably and we define them as infected in the Washington D.C. area just to see whether or not we can identify transmission networks as strikingly, so far Henry, we have been unable to do so.
So, in the analysis, about 120 different genotypes, we have been unable to identify significant transmission networks somewhat in contrast to what's been seen in other countries and that suggests that the transmission that works are so diverse that they can't be simply described as being an MSM epidemic or heterosexual epidemic or IV drug abuse epidemic. We don't have enough individuals in all of those groups to make that a very strong statement yet, but that's obviously something that we would like to continue.
GALLIN: Okay. Thank you very much.
It's now my pleasure to introduce our second speaker, Eric Engels, who will present recent finding in the epidemiology of cancer in HIV-infected people. Dr. Engels is a senior investigator in the infections and immunoepidemiology branch in the division of cancer epidemiology and genetics from the National Cancer Institute, he received his Ph.D. from Harvard and epidemiology in biostatistics at Tuft. After his internship and residency at the Women's Hospitals, Dr. Engels completed fellowships at the New England Medical Center in Boston. He came to the NIH in 1998 as a senior staff fellow in the viral epidemiology branch at the division of cancer epidemiology and genetics in the NCI, and his research concerns the role of infections, immunity and inflammation in a development of cancer. Particular interests include the association of chronic infections inflammation, and immunosuppression in the development of non-Hodgkins lymphoma and lung cancer.
Dr. Engels is a member of the American Association for Cancer Research, American Association for the Advancement of Science and currently he's on the editorial board for Cancer Epidemiology for Biomarkers, welcome.
ENGELS: Thank you, Dr. Gallin, and thank you for this opportunity to speak. It's a great pleasure to be here for you today.
Let me say I have no conflict of interest and I won't discuss off-label use of medications in my talk today. Here are the objectives of the talk. I hope we can understand the emerging trends and cancer risk in HIV infected people, understand the potential mechanisms related to HIV that are responsible to the development of these cancers and consider the research and clinical implications of these cancer trends.
So why are we interested in monitoring the epidemiology of cancer in HIV? Well HIV infected people are at increased risk for certain cancers. That's because they're immune suppressed and they frequently have infection with a number of viruses that themselves cause cancer, also people with HIV are exposed to other carcinogens such as those that are present in tobacco and alcohol. And an important feature of the epidemiology here is that since 1996, in developed countries we have had available a highly active antiretroviral therapy. And this therapy as I just discussed is effective at suppressing HIV replication, and it results in improved immunity and prolonged survival in people with HIV. And as a result, I’ll show you the incidence of a number of cancers has changed over time as a result of this therapy.
These observations have prevention implications, we need to pay attention to the cancers that are a major burden in this population, and further more understanding cancer in people with HIV, providing a window on the immune system in healthy people so it better helps us understand how healthy immune system protects all of us from the development of cancer.
So this figure, summarizes data from several studies it's a meta-analysis of studies of cancer risk in people with HIV, and on the vertical axis here is the standardize incidence ratio. Which is a relative risk, it compares the people of risk with HIV to the cancer in the general population. And you see these 3 cancers that are age defining cancers Hodgkin's sarcoma and cervical cancer, the risk are quite high especially for SRT sarcoma and non-Hodgkin’s lymphoma. These 5 cancers are some that I will discuss in detail today, these are lung cancer, Hodgkin’s lymphoma, anal cancer and liver cancer and melanoma cancer.
These increased frequency in people with HIV. In contrast, here are some common cancers that we see in the general population and these don't occur at increased frequency in HIV infected people. So it's really a specific pattern of cancer that we want to pay attention to in people with HIV.
Now a lot of the data that I’ll present today come from our national registry study, the HIV/AIDS cancer match study, a population based studies of cancer in HIV infected people. It's a computerized linkage of databases from U.S., HIV, AIDS and cancer registries. It includes 15 different geographic areas which are highlighted in red here. This study as data on 600,000 HIV infected people, including half a million people AIDS which represents about half of the entire U.S. total since the inception of the epidemic. And through this resource, we can evaluate the epidemiology of cancer in people with HIV for example, we can compare cancer risk to that in the general population. We can look at trends over time and we can study certain risk factors for cancer.
These are data from that study, this shows the changing pattern of cancer in people with AIDS in the united states, as a function of time, calendar year and what you see here for a KS, which is green, and for non-Hodgkin’s lymphoma in pink, then the incidence has declined dramatically over time. This is true even before 1996 before HAART became available, but this there was a decrease in 1996 as l. The non-cancers in purple, there incidence is pretty much stable, but now they represent a larger fraction of the total cancer burden and for that reason, if for no other, we need to pay more attention to these nonage-defining cancers.
Now, the first cancer I’ll talk about a little bit more is Kaposi sarcoma, it's an AIDS-defining cancer it's caused by herpes virus 8, and infected people, the risk increases as CD4 count declines and as more people become immune suppressed and HAART reduces the risk by 70-90%.
So these observations indicate the importance of driving development of this cancer and that's illustrated here using data from our registry linkage study, this shows the incidence of KS in people with AIDS in the united states as a function of CD4 count and you can see that it's a CD4 count declines, the risks for KS goes up, that's true in the early 1990s, but also in the period since 1996 when hard times become available although as you can see the risk dramatically declined with the availability of HAART.
Another point I would highlight is is a syndrome called KSiris immune resistance inflammatory syndrome. This has been described in recently started HAART. It is when people with HIV develop new KS lesions or there's dramatic worsening of KS lesions and it's thought that this worsening is due to infiltration of lesions by inflammatory cells and immune cells in the setting of therapy.
Now, the next cancer I’ll talk about is non-Hodgkin’s lymphomas. They're AIDS defining, there are 3 subtypes that are AIDS defining, diffuse large b-cell lymphoma and large central nervous system lymphoma and these are both tightly linked to the Epstein Barr infection and the third which is not an common is Burkett lymphoma now just with KS, the cancer increases as the CD4 count declines and the risk decreased with the available of HAART and again emphasizing immune suppression.
Here are data from Johns Hopkins Hospital from their large HIV clinic population there, and what this show that over calendar time, as you would expect there's been a dramatic increase in the use of HAART in the HIV population, so nowadays about 70% of people are treated with this combination therapy. In concert with that, there's been a decline in the incidence of KS. This curve actually shows the cumulative portion of people free of non-Hodgkin’s lymphoma, this is a curve for non-Hodgkin’s lymphoma and so before HAART was available, there is very steep.
So there are a lot of cases that developed but now in recent year it’s flattened off and that can be attributed to the use of more effective HIV therapy.
The next, the third aid to fighting cancer is cervical cancer this is caused by human papilloma virus and in women with HIV, having a low CD4 count is associate wide increased persistence of HPV and decreased clearance of the virus and it's also associate wide precursor lesions to cervical cancer, but it's been harder to show that the CD4 count is itself associate wide cancer risk itself and there has not been a decline in cervical cancer incidence with the introduction of HAART.
And that's illustrated here using data from our registry linkage study, can you see there's no clear association between CD4 count and the incidence of cervical cancer and there's really been no change over time and we interpret this to suggest that immunosuppression is important in the development of cervical cancer but it's probably only important at the earliest stages and once women have live wide HPV infection for a long time and have live wide HIV for a long time then the use of HAART to improve immunity doesn't top that progression and they can still develop cervical cancer. Now let me turn to the non-AIDS defining cancers. Here are the ones I mentioned earlier. All of these have established risk factors that are unrelated to HIV. These represent a varying fraction of the overall cancer burden in people with HIV. These are data from people with AIDS in the heart era and you can see that lung cancer is the most common of these non-AIDS cancers in people with AIDS, this represents 23% of all non-AIDS defining cancers. Almost a quarter and Hodgkin’s lymphoma also comprises a major subset of the non-AIDS cancers. All of these cancers have are occurring at increased frequency in people with HIV, compared to the general population, there really aren't as much data on skin cancers and I’ll talk about that later in my presentation. So let's turn first to lung cancer.
One of the major reasons why there's a lot of lung cancer in people with HIV, is that they frequently smoke. So here are data from a number of studies in the United States in people with HIV, illustrating that about anywhere from 40 to 70% of HIV infected people smoke and that has to do with the fact that the risk groups that are prone to develop HIV infection are populations where there's frequent tobacco use. But is that the whole story? Is it all just smoking?
Well, my colleagues and I have done several studies over the last few years and those are summarized briefly here to look at the epidemiology of lung cancer HIV. They used varying populations and varying designs, they were all large. This is the registry linkage study that I described to you and in each of these studies, there was an increased risk of lung cancer compared to the general population and in each of these studies we tried to adjust for the use of tobacco as the major risk factor for lung cancer and we use different methods and after those adjustments, there was still a higher risk of lung cancer in HIV infected people than in the general population. And let me illustrate that with one of those studies.
So this again, these are data from Johns Hopkins hospital from the HIV clinic, there were 33 lung cancers in that study of the HIV clinic population. That represented a 5 times the risk of lung cancer compared to the general population, significant excess and there was higher risk of lung cancer in males and females with HIV, and also for young adults and for older adults. Now does smoking explain that excess risk? Well, we knew from data on this cohort that almost 70% of these patients were current smokers, that's a very high proportion compared to for example, 23% of the Maryland general population and almost all of these lung cancer cases were in smokers. So tobacco appears to be a necessary factor. It's not like there's an epidemic of lung cancer HIV positive nonsmokers. But the amount that people smoked was not especially large. So in this clinic population, only 17% of people smoked more than a pack of cigarettes per day and almost a third smoked half a pack of cigarettes per day or less and among the people who actually developed lung cancer, they had smoked on average only 37 pack years at the time of their cancer diagnosis and that compared to 50 pack years in the HIV uninfected cases at Johns Hopkins.
So it looks like something is accelerating the development of lung cancer HIV infected smokers. And we accommodated for the effects of smoking on the risk of lung cancer in the statistical analysis indirect adjustment and here you see the sirs that I showed you before for lung cancer in red and then in yellow cancer in red and then in yellow the statistical adjustment and these sirs come down but they're all elevated. So this would suggest that even though smoke's important, it's not the whole story. And we've got additional analysis in our registry linkage study to look at the risk of lung cancer in people with AIDS as a function of CD4 count and those data are shown here, and you don't see a very clear relationship between lung cancer risk and CD4 count, so this is not your typical immune deficiency cancer like Kaposi sarcoma or non-Hodgkin’s lymphoma. And our working hypothesis now for what's going on in people with HIV, there are features of the disease that amplify the effects of tobacco use. For example, people with HIV have chronic lung inflammation, they get repeated lung infections and those may act synergistically with smoking to accelerate the developments of lung cancer.
The second cancer I’ll talk about is Hodgkin’s lymphoma. This is an EBV related tumor in people with HIV. The risk of Hodgkin’s lymphoma is including HIV patients but also transplant recipients and there are 2 paradoxical relationships that I would like to describe to you that are important. One is that we're now seeing that Hodgkin’s lymphoma risk is increased in people who are using HAART and further more and this is worrisome, the incident of Hodgkin’s lymphoma is increasing during the HAART era. That is illustrated here using a registry study, this looks at the risk of Hodgkin’s lymphoma in people with AIDS as a risk of calendar year and we don't have data for the most recent years on this figure so I wouldn't pay attention to those, but the point here is that during the 1990 when is there was mow effective HIV therapy available including the introduction of HAART, that the incidence of Hodgkin’s lymphoma rose and rose by a rate of 9% per year which is quite steep. And we think we might have a partial explanation for that and that's shown here, again using data from our registry linkage study, this shows the incidence of Hodgkin’s lymphoma as I function of CD4 count and as the CD4 count declines the risk of Hodgkin’s lymphoma increases as you might expect but then it peaks and drops off at the very lowest CD4 counts and now that we have effective effective HIV therapy, we think parents are being patients are being shifted from this part of the curve up to here where the lymphoma is higher and this may indicate that some cases of Hodgkin’s lymphoma are actually a type of immune reconstitution syndrome. Similar to what I was suggesting happens for Kaposi sarcoma.
The third cancer I’ll talk about, the third non-AIDS cancer and anal cancer, caused by human papilloma virus, particularly HPV 16, the risk is elevated in men who have sex with men but it's also high in other groups with HIV infection. And in people with HIV, having a low CD4 count is an associate wide increased risk for developing anal cancer precursor lesions but like for cervical cancer it's been harder to show that's the case for anal cancer itself.
Again I would like to highlight paradoxical observations, HAART is associate wide increased risk for anal cancer and people who appear to use HAART appear to have a higher risk of anal cancer and the incidence is rising over time. This is illustrated by data, from France from a large database in France, and it shows the incidence of anal cancer men who have sex with men, other males, females, all people with HIV and overtime, there's a rise in each of these groups in the occurrence of anal cancer and most of these bars are for the HAART itself, so what this data suggests is that the incidence is rising even within the HAART era.
Here are data from the U.S. study that look at the risk of anal cancer as a function of time, relative to AIDS onset. This x-axis is months since AIDS diagnosis and this is a measure of how long people have live wide HIV infection and we looked at it for in situ, and invasive cancers and in males and females separately and each of these categories, the risk increases the longer people have had HIV.
So what we think is going on in recent years that accounts for the increasing risk of anal cancer is that people have lived with HIV infection for a long time, they've had at least some modest immunosuppression for a long time and they've had hpv for a long time and at that point when HAART is introduced it's not effective at stopping the progression to anal cancer during those late stages or on the pathway to cancer. The other cancer that I’ll talk about is liver cancer. These are data on people with AIDS in the United States, here we this is from our registry study, we've broken the population down into groups where the (…) of hepatitis c is very high, hemophilia patients and injection drug users and groups where the prevalence is much lower. These are homosexual men and heterosexuals and other individuals and what you see here is that the risk of liver cancer is higher in the groups where hepatitis c is common, than in the groups where it's rare. And this means that this translates into an association with hepatitis c infection, so hepatitis c is responsible for some of these cases of liver cancer.
Another point is that the risk is high for liver cancer even in groups where hepatitis c is unanyone, the SIR is 5.5. That's a relative risk of 5 and half compared to the relative population. And this suggest there are other factors that are responsible as well, includes hepatitis b for example, and alcohol use.
Here are data from a recent Swiss case control study of liver cancer in people with HIV. One of the points I not to emphasize here is that all of these liver cancer cases in the Swiss study were infect wide hepatitis b or c or both and that emphasizes the importance of these viruses in causing liver cancer in HIV infected people. The other major finding in this study is that the risk of liver cancer increases if people have had a prior diagnosis of AIDS or if their CD4 count drops to a low level. So immunosuppression is also important presumably in allowing these viruses to escape immune control and that that's important in the development of liver cancer.
One finding in this study, final observation is that the association with low CD4 count was limited only to the hepatitis b infected subject so this might suggest that the importance of immune deficiency differs according to whether you're talking about hepatitis b related carcinogenesis or hepatitis c related carcinogenesis. Final group of cancers I’ll talk about are skin cancers here expose tower ultraviolet radiation is the risk factor and white vs the greatest risk because they lack protective skin pigmentation, there is a clearly increased risk for skin cancers in solid organ transplant recipients that's true for melanoma although the increased risk is modest in magnitude, it's also true for squamous cell carcinoma.
Here the relative risk appears to be much higher and patients transplant recipients can develop multiple highly invasive carcinoma lesions and finally, there are several rare cancers that are in current transplant recipients, these are mercocell carcinoma and sebaceous carcinoma which is a type of appendegeal carcinoma that arise from the support structures of the skin.
Notice my colleagues and I recently used our registry data to look at the epidemiology of skin cancer in people with AIDS, we couldn't look at basal cell carcinoma or squamous cell carcinoma because cancer registries don't capture that outcome but we did have a number of skin cancers. Melanoma was the most common of the types we evaluated. The risk was slightly elevated for this cancer, significantly elevated but not dramatically. More impressive are these rare cancers, merck cell carcinoma and sebaceous carcinoma the ones I mentioned where they're rare and the risk to the general population is quite substantial and for appendegeal carcinomaas as a group we were able to demonstrate that the rate increases with more time since an AIDS diagnosis, so the duration of HIV infection is a factor. And the immunosuppression and important in the development of these rare cancers of the skin. And in turn the fact that immunosuppression might be important suggests there might be virus there is that cause these cancer presidency in fact, there was recently described the detection of a new cell americale cell tumors.
Finally all we couldn't look at it, other people described that squamous cell carcinoma risk appears to be elevated in HIV patients. So taken together these observations would suck jest that skin cancers are under-recognized problem in people with HIV.
So let me conclude first by saying that I hope these data suggest a number of opportunities. I’ve tried to highlight some potential immune mechanisms related to HIV that could be important in the development of cancer. Those include inflammation and chronic infections. For example, those might be important in lung cancer. It may not be just the depth of the immune suppression but the duration of immune suppression that's important and I’ve tried to suggest that that could be the case for cervical cancer and anal cancer. And finally, there's the possibility that immune reconstitution in chromosomes could be important in some of these cancer as I’ve suggested for Kaposi sarcoma and Hodgkin’s lymphoma. As clinicians we need to do a better job of preventing cancer, I think we need to emphasize smoking cessation for HIV infected patients to prevent lung cancer and other heart and lung diseases and as well we can consider encouraging people to reduce their UV exposure by use protective closing and sunscreens. Those are similar recommendations that are advocated for transplant recipients.
And finally we can think about ways to screen and detect cancers early, there have been a number of arguments for a long time for anal pap smear screening to detect early anal cancers and we might also consider the use of chest x-ray screening or CT scanning screening to detect early lung cancers. I think I put a question mark here because I want to be tentative. This is an interesting idea but one that needed formal evaluation.
And finally there are reasons to be concerned. Although the cancer risk overall in people with HIV, has declined, that's really driven by the dramatic drops in the incidence of KS, and NHL, and there's have been burden cancers and for 2 of these cancers, the risk is increasing and those include as I highlight here Hodgkin’s lymphoma and anal cancer and finally, given the fact that people with HIV, are living longer and they're getting older. It's likely that the absolute risk of cancer in this population will continue to rise.
Thank you very much for your attention.
[ applause ]
GALLIN: Thanks to both speakers. It was great. Appreciate it.
[ applause ]
ANNOUNCER: We were pleased to present a pair of speakers on today’s edition of NIH Clinical Center Grand Rounds. Dr. Frank Maldarelli, from the HIV Drug Resistance Program at the National Cancer Institute discussed Genetic Variation of HIV-1 Populations in vivo: Implications for the Emergence of Antiretroviral Drug Resistance. He’ll was followed by Dr. Eric Engels, Senior Investigator in the Infections and Immunoepidemiology Branch of the Division of Cancer Epidemiology and Genetics, NCI, who presented Recent Findings in the Epidemiology of Cancer in HIV-infected People. You can see a closed-captioned videocast of this lecture by logging onto http://videocast.nih.gov -- click the "Past Events" link, or by clicking the "View Videocast" link for today's podcast at the Grand Rounds podcast page at www.cc.nih.gov/podcast/grandroundpodcasts.html. The NIH CLINICAL CENTER GRAND ROUNDS podcast is a presentation of the NIH Clinical Center, Office of Communications, Patient Recruitment and Public Liaison. For more information about clinical research going on every day at the NIH Clinical Center, log on to http://clinicalcenter.nih.gov. From America’s Clinical Research Hospital, this has been NIH CLINICAL CENTER GRAND ROUNDS. In Bethesda, Maryland, I’m Bill Schmalfeldt at the National Institutes of Health, an agency of the United States Department of Health and Human Services.
