Answers, Part Deux!

OK, that "answers" title is a bit misleading. I've only had time to answer one question this evening. So again, I'll ask you to be patient with me while I get to the next few.

Question 1: I've read a lot about the unreliability of the Eliza, Western Blot and Viral Load tests, including from Nobel prize-winner Kary Mullis. How do you account for this? Supposedly the proteins or dna/rna fragments tested for are present in normal blood anyway, and lots of things can cause false postivies?

Well, I’m not surprised that you’re a bit confused, given that so many people are deliberately attempting to sow confusion! Just so we’re all on the same page:

(1) ELISA (Enzyme-linked Immunosorbant Assay, or sometimes just EIA) + Western Blot: these test look for antibodies to HIV (not the virus itself, but the "criminal’s fingerprints")
(2) Viral Load (RNA): these tests (which are usually done by RT-PCR = reverse transcriiptase polymerase chain reaction, or sometimes bDNA = branched-chain DNA) look for HIV RNA in the blood (looks for "the criminal" himself).

For your diagnostic purposes, you want to be taking an antibody test - don't let anyone convince you that what you really need is a DNA PCR.

1. Antibody tests

when these are done properly, they are incredibly accurate (more accurate than the majority of diagnostic tests used in medicine). So, by asking whether the tests are reliable, the denialists appear to be “questioning” two things: (a) whether positive results on the ELISA + WB mean that a person has HIV (i.e. whether positive results are accurate); and (b) whether having HIV, as determined by antibody tests, means that a person will get AIDS (these people often argue in circles: "HIV doesn’t exist and is not the cause of AIDS, so the high accuracy of the HIV tests are meaningless because HIV does not exist and/or is not the cause of AIDS, and you can't show that it is, because all the tests are meaningless!"). Duesberg and Mullis hold that the tests are accurate, but that it doesn’t matter because HIV doesn’t cause AIDS. Other denialists maintain that the tests are inaccurate to diagnose HIV and/or are COMPLETE bull because HIV does not even exist. Here, I'm going to try to address the accuracy issue, because you asked a separate question about the Duesberg/Perth arguments which I probably won't be able to get to today, but some of that stuff will inevitably come up.

So, HIV diagnoses should be made on the basis of a positive ELISA, confirmed with a positive Western Blot. The tests work in conjunction with each other: The screening is like this for a reason. You don’t want to miss people with HIV, but you also don’t want to give people incorrect HIV diagnoses. The ELISA is extraordinarily sensitive: job is to pick up people who probably have HIV and make sure that they don’t slip through the cracks (hence, you might get some false positives; it's a cheap, easy screening tool). Because of its sensitivity, diagnosis is not based on a positive ELISA alone. It has to be confirmed with a Western Blot. The Western Blot has a higher specificity (lower rate of false positives).
You can think of it like this: the ELISA is like a cheap burglar alarm that can occasionally go off when a burglar is not in the house (i.e. it picks up activities that seem like a burglar, because it would rather have a “false positive” than a “false negative”), the Western Blot is a more sophisticated, expensive motion detector or video monitoring system that is better at telling whether the thing that set off the burglar alarm is actually a burglar. ELISA rules in people who MIGHT be HIV-positive, Western Blot rules out those who aren’t HIV-positive. In practice, it is difficult to increase sensitivity without decreasing specificity. This is why we have a two-step protocol.

How do we know that HIV-testing is accurate?
Because it has been validated! So, if you've just invented a new screening test for HIV, how do you prove that it is accurate? Well, you would take people whom you can prove have HIV in some way other than the tool you’re trying to validate (e.g. those from whom you can culture the virus), test them, and then test people from whom virus can’t be cultured. If the test is bad, the results will not match the culture. If the thing you’re testing for (anti-HIV antibodies) is something that you will find in any junkie, gay man, or hemophiliac, then those hemophiliacs, junkies, and gay men from whom HIV cannot be cultured will nevertheless test positive.

These tests have been extensively validated, much more so than other antibody tests used in medical diagnostics. The reason for this is pretty obvious: HIV was (and, perhaps to a lesser extent, continues to be) a “dread diagnosis” and given the psychological devastation that a false-positive result could inflict, it would be profoundly unkind to tell hordes of HIV-negative people that they have this virus. Also, if you are trying to prevent the spread of this virus, you want people who are HIV-positive to know their status so that they won’t unknowingly infect others -you therefore don’t want a high number of false-negatives either!

If you are looking for a perfect test, you are unlikely to find one. A perfect test would be one that was 100% correct all of the time, would not be subject to laboratory error, and could give a positive result even if the person was infected a nanosecond before the test was taken. This is unlikely to ever happen.
The HIV test is not perfect (no test used in the field of infectious diseases is perfect). But it is pretty damn good:

http://www.ncbi.nlm.nih.gov/pubmed/2298875?dopt=Abstract
(409 people who tested positive by Western Blot and 131 people who tested negative. Validated by culture and DNA PCR – the virus was found in 100% of the people who were antibody-positive, and 0% of the people who were antibody-negative).

This was a small sample (n = 540). It’s undoubtedly true that in larger samples, some false positives do occur. But how often?


http://www.ncbi.nlm.nih.gov/pubmed/3419477?
(study on the military’s testing of applicants between 1985-1987 - when antibody tests were less accurate than they are now, by the way! False positives were 1 in 135,187)

http://www.ncbi.nlm.nih.gov/pubmed/9757856?
(1998 study on false-positives among blood donors: false positive rate was 1 in 251,000 based on RNA PCR and follow-up serology. That’s .0004% of donors, or 4 per million).

If they believed what they were saying, the denialists who are advocating “STOP HIV-TESTING NOW!” because false positives exist should be advocating for the end of most types of medical testing completely. If there were not an HIV-test (and in the early 80s, there wasn’t one), a hell of a lot more people would be at risk of incorrect diagnoses. The majority of Americans who get tested are HIV-negative and receive negative HIV tests – nobody would able to reassure these people that they were not in fact infected (right now, you’re experiencing the anxiety of “not knowing” – how would it feel to you if there were no way anyone could ever tell you that you were negative, or if the only way you could really know you were positive would be when you actually developed AIDS?). Members of groups who are “at high risk” of HIV (e.g. people have engaged in risky sexual behaviors such as receptive anal sex or sex with people from countries with high HIV prevalence rates) would have to assume that they were positive, even though many would not be. Doctors would do repeated CD4 counts and CD4:8 ratios on asymptomatic individuals in an attempt to see “a decline”, and this would lead to a lot more anxiety and false-positives than the HIV test ever could. Everyone would be paranoid, and yet the virus would probably spread like wildfire!

In arguing that HIV tests are inaccurate and have unacceptably high rates of false positives, denialists construct strawmen. You can decide for yourself whether we should stop HIV-testing based on false-positive rates of 1-8 per million, but you should understand what their arguments are. Instead of tackling the issue (and data!) head-on, denialists construct theoretical arguments of why there COULD be false positives. So, I guess I should start by making sure you know what a positive test means: it means that you have IgG antibodies (usually IgG - some of the newer ELISAs will also look for IgM and/or IgA, but as far as I know, the commercial Western Blot only looks for IgG) that bind to (a) ANY part of HIV (ELISA) and (b) multiple specific HIV antigens (Western Blot). Basics: an antigen is something the immune system recognizes and responds to - pathogens are antigenic, but so are plenty of harmless things. The pieces of antigens that induce specific B-cell and T-cell responses are called “epitopes”. IgG – immunoglobulin G, sometimes called gamma-globulin – is a type of antibody that is very prevalent in the blood; IgM, Immunoglobulin M, is the first type of antibody a B-cell will make, so it often appears earlier in infection; IgA, Immunoglobulin A, is another antibody type that is very prevalent in mucosa and in secretions.

So when some nasty pathogen (say, smallpox) enters your body, it will eventually get to your B-cells. You make millions of B-cells and T-cells every day, each one recognizing a different epitope. Generally speaking, B-cells and T-cells are “antigen-specific”, which means that a B-cell that recognizes a staphylococcus antigen will not recognize an ebola virus antigen. This isn’t a great analogy, but it’s as if you were making millions of keys every day, so when a given lock came along, you would most likely have a key to fit in it. Most of these B-cells (keys) will die off without ever being used (being activated). However, when you are infected with smallpox, it eventually gets to the B-cells (I’m skipping a lot here), and a few B-cells will recognize it: their keys (receptors) will fit the smallpox locks (antigens; pathogens have lots of potential antigens, so more than one B-cell will recognize it, but they will be recognizing different pieces of it). Once a B-cell meets its antigen, it is activated, and a bunch of other stuff happens and then it starts secreting antibodies. A smallpox B-cell secretes antibodies to smallpox, an ebola virus B-cell secretes antibodies to ebola virus – for these purposes, the antibodies fit in the lock the same way the B-cell did.

Now that you know this, a few of the reasons that false positive results can occur might make sense to you. A B-cell specific to some piece of HIV will generally not start secreting antibodies until it has come into contact with that HIV antigen and been activated by it. This is why antibody testing usually works: because you do not have antibodies to things that you have not come into contact with; if you test positive on a test for antibodies to Hepatitis C Virus, that means you have come into contact with the Hepatitis C antigen they are testing. But there are wrinkles to this. Since the immune system does not recognize “whole pathogens” but rather pieces of pathogens, if two pathogens have similar pieces, you can get what is called “cross-reactivity” (for example, a T-cell that has responded to tuberculosis may also react weakly to closely related mycobacteria like leprosy or M. avium).

The ELISA works by taking some of your blood (in which there are all sorts of antibodies! If you were vaccinated as a child, you’ve got antibodies to measles and rubella and polio and tetanus everything else [unless you haven’t kept up with your boosters, you bad boy!]. Not to mention all the antibodies you’ve picked up by being exposed to foreign antigens all your life) and showing your blood some purified HIV (the whole virus) and seeing if antibodies (usually IgG) bind to it. Steps are taken to prevent non-specific binding by other antibodies (non-specific binding sites are blocked by proteins – it is like putting tape over a lock that you’re worried other people’s keys may be able to fit into; your blood is also diluted in a way that further prevents non-specific binding. And then when your blood is added, any HIV antibodies that you have will bind to the purified virus.

Antibodies are a bit like magnets: if they’re specific to a germ, they’ll stick to it! The antibodies that aren’t specific to it can only stick to it accidentally; they’re not very sticky, so when you wash the virus, everything except the antibodies that are specific to it should be washed away. It’s like if you had a refrigerator lying face-up on the ground and some bags of all sorts of junk including some magnets. If you were to drop the contents of those bags onto the refrigerator, all sorts of stuff, including the magnets, would wind up on the refrigerator’s surface. But then if you took a firehose and sprayed it at the refrigerator for a while, the superfluous junk would be washed away and only the magnets would be left. If denialists are talking about proteins other than antibodies causing a false-positive, they don’t know what they’re talking about. You could have a gazillion proteins sticking to the virus used for the test, and they wouldn’t cause a false-positive – the only thing that can cause a false positive is the type of antibody you are looking for (e.g. if you are looking for IgG specific to HIV, the only thing that can trigger a positive result – besides laboratory error or mislabeling – would be IgG. I won’t go too deep into how we’re able to “see” that the proteins on the virus are IgG, but it’s a bit like if the magnets that attached to the virus on one side had velcro on the other side - now we can throw a bunch of velcro that we marked with bright paint onto the refrigerator - the velcro will only stick to the velcro-covered magnets (not to the refrigerator or any other crap that may be left there; pretend like velcro is attracted to other velcro the way magnets are). Now we shoot that firehose again, to wash away any painted velcro that is not stuck onto the magnets. This velcro will not attach to random proteins, it is designed to stick only to a particular region of IgG).

So, if you are talking about a false positive, what you are essentially talking about is a person without HIV nevertheless having some antibody that is binding to HIV. With the ELISA, this is entirely possible (although still not all that common). There is a cut-off point, above which the number of “velcros” seen is considered to be positive (this is because some non-HIV antibodies may bind or stick even after washing, but there usually won't be nearly as many of them as there are HIV-specific antibodies – they’re just not “magnetic” enough to be attracted to the HIV with the same avidity). If you have more antibodies in your blood than “normal”, there is a small chance of a positive ELISA (so this may happen following certain infections or vaccines, with some antibody-mediated autoimmune diseases, with certain alloantibodies following a graft or pregnancy, or with medical conditions like hypergammaglobulinemia, which is just a fancy word for ‘too many IgG proteins in the blood’)

This is one reason why reactive ELISAs are confirmed with the Western Blot (you can’t diagnose HIV on the basis of an ELISA alone. Actually, before the Western Blot is done, the ELISA is usually run a second time, in an attempt to make sure that the positive result was not due to laboratory error). On the ELISA, the whole virus is used. Without getting too technical, on the Western Blot, you are using individual HIV antigens. If you expose a person to HIV, they will make more than one kind of antibody, because there is more than one antigen on the virus. So, people who are actually HIV-positive will have different types of antibodies to many different pieces of the virus. You can see why a test for antibodies to these pieces is more specific than a test for antibodies to the whole virus: a person with “too many” antibodies or cross-reacting antibodies may have some antibody that sticks to some piece of the virus, but it’s very unlikely that they will have antibodies that cross-react with more than one HIV epitope in the particular pattern required. A Western Blot is not interpreted as positive just because there is some antibody responding to any old HIV protein. To have a positive Western Blot, you need to have multiple anti-HIV antibodies reacting to very specific HIV proteins (I’m not going to get into the specific patterns that indicate a positive result unless someone asks, since if you’re not asleep yet, you would be soon!). If the antibodies react in the right pattern, the person is diagnosed HIV-positive. If the person’s antibodies don’t react with any of the HIV epitopes, the test is negative.

But there is another possibility: the person has antibodies that react to one or more proteins, but not the pattern of proteins required for a positive result. If it’s not the right pattern, the test should NOT be considered positive. It’s “indeterminate”. One reason that it’s indeterminate (rather than simply negative) is because while people are seroconverting (i.e. in the early weeks following infection), they will often get indeterminate results because they are making some antibodies to HIV, but they are not yet having the antibody response that typifies chronic infection. However, not everybody who gets an indeterminate result is in the early stages of infection, so everyone with an indeterminate result needs to be re-tested after a period of time (different people recommend different times, but it’s at least one month. Some doctors may run a PCR in the meantime). Indeterminate results are, in the scheme of things, not that common either (although they’re more common than false positives. You’ll hear denialists say that HIV results are “always positive” or “completely inaccurate” during pregnancy. That’s incorrect. They’re taking a kernel of truth – that indeterminate results occur more often during pregnancy than at other times – and twisting it). There is a lot of confusion among laymen about what an indeterminate result means, and the denialists play on this when they talk about indeterminate results as if they were false-positives (they’re not! A doctor once diagnosed someone as HIV-positive on the basis of an indeterminate result and he got his ass sued big-time because nobody with an indeterminate result should ever believe that their tests showed them to be HIV-positive). Such results can obviously be psychologically difficult – the person may have to wait a long time until they know anything for sure, and it can create a lot of anxiety. However, it is necessary to do it. The next test will show them either to be positive (in which case they were probably seroconverting), negative (in which case the first one was a fluke), or indeterminate. Repeated indeterminate results do happen, and they should be taken to mean that the person is HIV-negative.

Do true “false positives” (as opposed to indeterminates, which should not be interpreted as positives, false or otherwise) happen? Yes, extraordinarily rarely (4-8 times per million). Laboratory error - usually a mislabeled specimen - is always a possibility. False positives can occur in people who have participated in clinical trials to develop HIV vaccines (since the vaccine triggered the same antibody response that they would have if they were infected). The only medical condition I know of that creates a significant risk of false-positives (as opposed to indeterminates) is lupus. Lupus is an autoimmune disease where people make antibodies to their own DNA (this is a bad thing for the obvious reason: you don’t want your immune system attacking your own cells!). There are probably pieces of our DNA that one could make antibodies to that are SIMILAR to certain pieces of HIV (NOT the same – I'll probably talk about this when I get to viral load), and they are similar enough that they can cross-react.

Here's a good study on the accuracy of the two-step (ELISA + Western Blot) protocol:

http://www.ncbi.nlm.nih.gov/pubmed/2648922

630,190 blood donations from 290,110 donors in Minnesota over a three-year period in the 1980s. All donors who tested positive on the antibody tests were tracked down (except for one, who couldn't be located). They ran cultures, and every single one was positive for HIV. No false positives. They put the false positive rate of this protocol at between 0 and 6 per million.

As you can see, the protocol for diagnosing a person HIV-positive is somewhat complicated and can be repetitive. The low false-positive rate speaks to how well this works the vast majority of the time. When false-positives occur, they are hopefully caught fairly early by other tests conducted by an experienced practitioner. Nevertheless, there’s no doubt that false-positives can be traumatic for the small number of individuals who are affected. Without denying the impact that a false-positive can have on someone, I don’t think any reasonable person could suggest that the very low risk of a false-positive outweighs the enormous benefits that comes from testing. The denialists are simply lying if they claim that nobody is paying attention to the risk of false positives (the number of papers written on the subject, not to mention all the government efforts to validate testing methods and determine proper protocols, speaks to the importance that mainstream Science attaches to accuracy of results). Because accurate HIV-testing is so important to everyone who wants to curb this pandemic, the HIV test is truly one of the best diagnostic tests in infectious diseases. Anyone who has a problem with HIV testing has a problem with medical diagnostics, and perhaps the entire field of modern biology. If the denialists believed what they said, they would be railing against the use of EVERY antibody-based test (ELISAs and Western Blots are exceedingly common in the diagnosis of infectious diseases. It’s unclear to me why the denialists are not simultaneously maintaining that Hepatitis B Virus does not exist and/or does not cause cirrhosis or hepatocellular carcinoma – the latter would be more plausible than maintaining that HIV doesn’t cause AIDS, although it still wouldn’t be very plausible!).

What’s particularly ironic is that there are a few denialists who are financially involved in the alternative health sector where they sell all sorts of remedies for all sorts of conditions, real and imagined, which can’t be diagnosed with nearly the precision of HIV. Years ago, I met an HIV-positive denialist who blamed his deteriorating health on “candida”, a condition that mainstream medicine does not acknowledge (as opposed to candidiasis) and for which there is no valid test. The logic of rejecting an HIV diagnosis as “speculative” but accepting a candida diagnosis as "valid" is simply beyond me!

2. Viral Load
So, my friend, if you are still with me, now we turn to VIRAL LOAD!
To oversimplify it a bit, viral load refers to a test that counts the number of viral particles (HIV virions) per unit of blood. There is another test, DNA PCR, that counts the number of virally-infected cells per unit of blood. As you can see from the names, the test looking for virions looks for RNA; the test looking for infected cells looks for DNA. HIV is an RNA virus (which means that its genetic material is contained in the form of RNA; this is not true of all viruses – DNA viruses, like the herpesviruses, keep their genetic material in the form of DNA, which is how plants, animals, fungi, and bacteria keep their genetic material).

So, if you are looking directly at an HIV virion and want to see its genetic material, what you want to see is RNA. However, HIV is a retrovirus! If you’ll remember 9th grade biology, in living organisms (viruses aren’t really considered ‘alive’, although there’s some debate), our genetic code is kept as DNA, transcribed into RNA, and translated into proteins. So transcriiption is normally DNA into RNA. But retroviruses can reverse transcribe! They can take their RNA and transcribe it into DNA, with the help of an enzyme called ‘reverse transcriiptase’ (this is one of the targets for two of the HIV drug classes. HIV needs reverse transcriiptase to breed new virions, and human RNA does not reverse transcribe, so if you inhibit reverse transcriiptase, you can inhibit the virus without messing ourselves up too badly). HIV only reverses transcribes AFTER it has infected a cell: HIV attaches to the cell, dumps its genetic material and enzymes in the cytoplasm, and then reverse transcribes so its RNA is now double-stranded DNA. It’s like going from VHS to DVD – same material, different form, but the change in form is necessary because the “translation machine” in a human cell’s nucleus only plays DVD. Once the HIV has reverse transcribed to dsDNA, it heads over to nucleus where it integrates into our DNA (it basically cuts a hole in our DNA with a viral enzyme, and then fills in the gap). Once HIV is part of a cell’s DNA, it can use the cell’s machinery and use it to make more copies of itself (again, I’m trying to make this as non-technical as possible, but if anyone wants more details, I’m happy to give them). So, you see now why if you are looking for virions outside the cell, you are looking for RNA; if you are looking for infected cells, you can look for DNA.

The viral load test is the RNA test – it’s looking for virions, because they’re a better gauge of active replication (the half-life of HIV virions is only 8 hours, so if you have a lot of virions, you know the virus is replicating at a high level. By contrast, since HIV can be latent – dormant – once it is inside a cell, the number of infected cells does not necessarily tell you how much the virus is replicating right now. The real difficulty in curing HIV is in deleting those cells where HIV has already become part of our DNA. Right now, we can stop replication, but we can not remove the HIV that has become part of our own genetic material).

How do we look for HIV virions and count the number of copies? There are multiple techniques, but the most common one is PCR (Kary Mullis’ brainchild), followed by bDNA. Denialists usually attack PCR and ignore bDNA, because bDNA is a direct method of quantifying the virus, while PCR is an indirect one. It’s pretty obvious that if you want to “see” the genetic material from a virus, you have to “magnify” it somehow (you’d have to take all the blood from a whole lot of people’s bodies to get enough HIV RNA to see it without magnification, or even with a very good microscope). You also need some way of quantifying it – even if it were possible to individually count every copy of the genome, viral loads can be in the millions, and no lab technician is going to waste their time on that (and if they did, the likelihood of inaccuracy would be huge. Humans aren’t very good at counting to high numbers without getting distracted and forgetting their place – starting over is often such a pain that people will just guess where they were).
A person’s viral load is not actually a precise count, it’s just a very good estimate. Here too, denialists create strawmen: they’ll say “viral load tests can’t do XYZ”, when nobody said that they could! A viral load test gives a good estimate of the amount of viral nucleic acid in the blood, and therefore the number of virions in the blood. It does NOT tell you how infectious those virions are (although even non-infectious virions can cause a lot of problems!), whether viral load is higher or lower in other parts of the body, etc.

I’ll explain bDNA first, since it’s a bit easier to understand: You have some blood, you want to see if you can detect HIV nucleic acid in it, and if so, how many copies. After some initial steps that basically work to “get the RNA where you want it”, you still aren’t going to be able to literally SEE the RNA. You need to make it “stand out”. To oversimplify it with an admittedly imperfect analogy, it’s like you have a bunch of magnets that can each shine a certain amount of light – you can draw a genetic sequence on the bottom of the magnets so that they will only stick to that particular genetic sequence; once a magnet sticks to its genetic sequence, the light turns on. After you do that, you can then measure the amount of light they are emitting. Then, you can figure out how many magnets (copies) it takes to emit that amount of light, and that number is your viral load.

With RT-PCR (reverse transcriiptase polymerase chain reaction), instead of making the signal brighter, you are making the amount of nucleic acid bigger with a “chain reaction”. Again, it is like you draw a genetic sequence on a bunch of magnets, and send them in. But instead of emitting light, when the magnet sticks to the genetic sequence you are looking for, it makes a copy of it (a cDNA copy, just because that’s easier to work with). Again and again, you drop more round of magnets on it, and each time the sequence is copied, until you have about a million times as many copies of that sequence (as DNA) as you did before (I’m skipping some of the details here, because it’s hard to make a coherent analogy out of this, and I think you just want to understand the point, not know all the steps!). The magnet is not only copying the sequence, it is also adding a chemical so that when the chain reaction is complete, you can add another chemical to the solution which will make the first chemical change color. You can then measure how intense the color is (by machine, of course!), and use that to calculate how many DNA copies there are now, and by extension, how many RNA copies there were originally. DNA PCR is basically the same process, except that you are looking for a DNA sequence instead of an RNA one.

Do either of these tests give you a precise count of how many viral particles you have in your peripheral blood? No! They give good estimates that are, more or less, internally consistent. They are a tool, and a very useful one at that. PCR results do have to be used (and interpreted) properly. If your viral load goes from 10,000 to 100,000, does that mean that every ml of your blood has exactly 90,000 more copies than it did before? Of course not! What it does mean is that the burden on your body from the virus is increasing, your immune system is losing the ability to keep its replication in check, and you are at a greater risk of disease progression and death. A high viral load, alone, will not make you prone to opportunistic infections (although the viral particles can do damage in other ways) – if you have a high viral load and a high CD4 count, you would be unlikely to get opportunistic infections at the present time. But, the high viral load predicts opportunistic infections in the future. A rough analogy that is sometimes used is that you are on a train track, with the stationhouse at the end of the track being serious opportunistic infection or death. Your CD4 count tells you how far the train is from that stationhouse, the viral load tells you how quickly you are moving. Is there individual variation? Of course! Tons of it! Nobody without a crystal ball will tell you exactly when you will progress to AIDS, but the viral load helps give you a reasonable estimate of your risk. Viral load testing is also invaluable in monitoring the efficacy of HAART, since if a person has a detectable viral load on HAART, his virus is able to replicate, and there’s a much higher chance of treatment failure and/or the development of drug resistance (the fact that most people on HAART are able to achieve an undetectable viral loads lays the lie to the claim that the RNA sequences tested for by PCR could be human sequences that all of us have).

Since I brought up Mullis, I’ll just note here that denialists often use old quotes from Mullis about HIV being an improper use of PCR (HIV was the first disease for which PCR monitoring was used on a widespread basis. Since then, PCRs have been developed to detect a whole host of other infectious diseases, and Mullis’ theoretical objection – that the technique is just too sensitive and not specific enough – would apply to all of them), but Mullis himself seems to no longer believe this (although he still believes that HIV is not the cause the AIDS). Last year, an HIV+ Australian man was prosecuted for intentionally infecting others (which is different than unknowingly or negligently infecting others. There have been some spurious prosecutions along these lines in this country, but the issue has always been ‘intent to transmit HIV’ rather than ‘does HIV cause AIDS?’) and he attempted to defend himself (with Perth Group support) by saying that HIV didn’t cause AIDS. They tried to cite Mullis’ early comments on the unreliability of PCR in this context, and one of the prosecution lawyers contacted Mullis for comment. His response was:
“I will not try to convince anyone that PCR can be used successfully to specifically make multiple copies of any nucleic acid sequence that can be uniquely defined by two “primer target sequences” comprising the termini of the sequence of interest. The veracity of this no longer has anything to do with me. I think this has been confirmed by a huge number of laboratories around the world. The rapid spread of this simple technology would not have occurred had it been ineffectual or flawed in any persistent way.
The matter which you are considering, if I understand it correctly, is that the presence or absence of a given nucleic acid sequence, as determined by PCR, can be used as a reliable marker for a living organism in a biological sample. This is done quite often in scientific studies, but that does not mean there could never be exceptions. Remember scientific studies are done with the understanding that findings will be subject to scrutiny from colleagues. A nucleic acid segment very similar in size and terminal base could easily, in a cursory examination, be mistaken for the sequence in question. If this happened in the course of a normal scientific finding, somebody would finally notice it.”

His belief is NOT that PCR quantification of HIV is inaccurate, but rather that HIV has nothing to do with AIDS (like Duesberg). You may have noticed that many denialist websites offer huge cash prizes to anyone who can present a paper showing that HIV has been isolated. You know who has tried to claim these prizes? DUESBERG! Here is what Duesberg has had to say when he tried to collect a 50K prize (I’m not sure what it is called now. Originally it was the “Continuum Challenge”, named after the denialist publication Continuum, which went under when every single one of its editors died of AIDS. I can’t keep the ever-changing names and groups straight because I really don’t pay all that much attention to them, but it's obviously whatever prize has been offered by the Perth Group). Duesberg's words:
“In conclusion: HIV has been isolated by the most rigorous method Science has to offer. An infectious DNA of 9.15 kilo bases (kb) has been cloned from the cells of HIV-antibody-positive persons, that -upon transfection- induces the synthesis of an unique retrovirus. This DNA "isolates" HIV from all cellular molecules, even from viral proteins and RNA. Having cloned infectious DNA of HIV is as much isolation of HIV as one can possibly get, it is like isolating the fifth symphony from an orchestra hall by recording it on a CD. The retrovirus encoded by this infectious DNA reacts with the same antibodies that crossreact with Montagnier"s global HIV standard, produced by immortal cell lines in many labs and companies around the world for the HIV-test. This confirms the existence of the retrovirus HIV.
The uniqueness of HIV is confirmed by the detection of HIV-specific DNA sequences in the DNA of most antibody-positive people. The same DNA is not found in uninfected humans, and the probability to find such a sequence in any DNA sample is 1/49500 - which is much less likely than to encounter the same water molecule twice by swimming in the Pacific ocean every day of your life” (you can read the rest on Duesberg's website here, if you care to:
http://www.duesberg.com/papers/continu1.html)


Notice that Mullis and Duesberg, the two acclaimed scientists that denialists claim as their own, both admit that HIV exists, can be isolated, and are reliably detected by tests like PCR (and antibody tests that detect antibodies against them). This is probably because they, unlike the majority of denialists, understand how the technologies actually work. Notice how Mullis says that it is theoretically possible that PCR could detect the wrong thing (“could never be subject to exceptions”; “a nucleic acid segment very similar in size and terminal base could easily, in a cursory examination, be mistaken for the sequence in question”), but also says that this is the sort of thing that SOMEBODY WOULD NOTICE! Scientists make errors all of the time, but Science is, over time, self-correcting. Denialists like to paint modern scientists as a bunch of sheep who refuse to challenge current “dogma”. I don’t think they’ve really spent much time with research scientists! Every scientist wants to make a really important discovery that makes a name for himself. Science often proceeds by drips and drabs: it can be an awful lot of work to make a small discovery that contributes something important to the overall picture, but doesn’t really change the picture drastically. If his work is valid and he does it well, he’ll have respect within the scientific community, but that’s a far cry from fame and fortune. BIG discoveries that change current conceptions and models garner a scientist much more attention than work on something like alterations in CXCR3+ CD8 T-cells in advanced infection. Scientists in thousands of labs worldwide use immunoassays and PCRs – if they didn’t work on HIV, if it turned out that the things we are testing for exist in all people, that people without the virus have anti-HIV antibodies and cross-reacting nucleic acids, that would be a big deal! They would be clamoring to publish it! Anyone who claims that there are “secret” discoveries that have been “suppressed” just does not understand how scientists are – they’re like piranhas when it comes to publishable data! No scientist is going to say, “yeah, I’ve made this really important and unexpected discovery, but I’m willing to give up the fame and fortune that could come with it, because I care neither about advancing scientific knowledge nor about my own status within the scientific community. You’re right, I should just let it go!”

If the utility of viral load testing was ever in doubt, the fact that it has shown great predictive value should put that to rest. Contrary to what many claim, HIV has been cultured, time and time again, from AIDS patients. They claim that the cultures are invalid because in order to keep the virus alive long enough to culture it, you have to “feed” it more cells (viruses are obligate pathogens; in order to replicate, they need to take over host cell machinery. You can often culture bacteria on cell-free media like agar, but for viruses, you need some kind of living cell that they're able to replicate in. The problem was that before the virus could be cultured, it would kill off all the T-cells in the culture, and without any cells, you’re not going to be able to grow more virus. So you had to feed in new cells to replace the ones that were dying off) – the denialists’ claim is that since new cells were added, the virus that was isolated was merely a contaminant that came from those cells and therefore did not come from the AIDS patients bodies. If this was ever a plausible claim (the same cell lines had been used for all sorts of reasons in the past, and this ‘contamination’ never occurred), it should not be one now. The internal consistency in our understanding of HIV is overwhelming. If the virus cultured from AIDS patients was a mere contaminant that did not come from their bodies, you should not be able to find the genetic material of that contaminant in their bodies – and yet, you can (that’s exactly what PCR does! See
http://www.springerlink.com/content/g1541k1rg0732118/
- qualitative DNA PCR; 100% of HIV-positive patients had positive results by PCR; 100% of HIV-negative donors, including those who had indeterminate results on the Western Blot, had negative results). If AIDS patients did not have that “contaminant” in their bodies, why do they have specific antibodies that react to more than one of its proteins, while those in whom you cannot find this “contaminant” do not? Why do the RNA levels of this “contaminant” in their bodies correlate so well with the risk of serious infection and death in populations (i.e. an HIV+ person with a viral load > 30,000 has 18.5 times the risk of death as an HIV+ person with a viral load <500)? You can call it human immunodeficiency virus or you can call it the creature from the black lagoon! The name is irrelevant – what’s relevant is that it will kill you!

I’m not going to talk too much more about the different uses for PCR, because what you really need right now is NOT a PCR, but an HIV test (ELISA and, if positive, Western Blot). It’s been a couple years since your exposure, so the only justifiable use for a PCR for quasi-diagnostic purposes in adults has passed. Using a PCR for diagnostic purposes isn’t what it’s cracked up to be (hence, it isn’t recommended!). It’s much more expensive, more time consuming for the lab, and more easily screwed up. You also can’t do it anonymously, because your doctor has to order it (if you look in other forums, there are tons of people who, despite having no real risk for being HIV-infected, worry about it endlessly. They take the HIV test 5 times, and when the results are negative, bully some uninformed doctor into running a PCR. Of course, the results are negative!). An antibody test is also just a better diagnostic instrument in most cases.

I probably wrote too much, and yet, I could have written so much more! So, I’ll just send what I’ve got, and come back to certain issues when I answer your next question about denialism. Now you know how these tests work (and why they are very accurate, and yet have limitations). Any questions or confusion, just ask!
Tweet