HIV vaccine shows promise

September 24, 2009 8:48 AM

A new HIV vaccine is showing promising results, reducing the risk of contracting the virus by 32 percent. While further tests are still needed, the vaccine is a combination failed HIV vaccines AIDSVAX and ALVAC, based on the Canary Pox virus. The study itself faced criticism from the outset.
posted by borkencode (41 comments total)

32% sounds better than nothing, but considering the illusion of immunity it could perpetuate among those who are vaccinated, I'm wondering if it might not actually be better than nothing.
posted by hermitosis at 8:52 AM on September 24, 2009

In case you suddenly got a little scared, Canary Pox is only for canaries.
posted by smackfu at 8:54 AM on September 24, 2009

When there is a (near) 100% effective AIDS vaccine/cure there is going to be such a public orgy.
posted by DU at 8:55 AM on September 24, 2009 [8 favorites]

32% sounds better than nothing, but considering the illusion of immunity it could perpetuate among those who are vaccinated, I'm wondering if it might not actually be better than nothing.

I assume you meant 'might not actually be worse than nothing.' As some of the coverage points out, a risk reduction of 50% is considered the threshold of "unequivocal clinical benefit," so the vaccine will not be deployed without further refinement.
posted by jedicus at 8:55 AM on September 24, 2009 [1 favorite]

32% sounds better than nothing, but considering the illusion of immunity it could perpetuate among those who are vaccinated, I'm wondering if it might not actually be better than nothing.

I suppose that depends on how the vaccine is distributed. I’m guessing any implementation would involve lots of education and lots of condoms. %32 sounds pretty stellar to me, but I have no idea what kind of success rates these studies normally shoot for.
posted by Think_Long at 8:56 AM on September 24, 2009

I'm not a scientist or anything here, but 32% just sounds...odd. Like the lack of infection may have been more to do with safe sex practices, etc., than the vaccine itself. I would think that a working vaccine would have a much higher rate of prevention. I would be happy to have come to the wrong conclusion, obviously.
posted by kittens for breakfast at 9:04 AM on September 24, 2009 [1 favorite]

THERE IS NO CURE FOR CANARY POX!!! Canaries are doomed! The coalmine is earth! We are doomed!

Wait, what? Oh HIV. Sorry. Back to topic. From the first link:

The vaccine lowered the risk of HIV infection by 32 percent among 16,000 heterosexual Thai volunteers who had no special risk of AIDS infection, the U.S. and Thai government researchers said.

Two questions: 1) why the need to clarify that the volunteers were heterosexual? and 2) how can you test for a lowered risk of infection without attempting to infect someone?
posted by filthy light thief at 9:12 AM on September 24, 2009 [1 favorite]

32% sounds better than nothing, but considering the illusion of immunity it could perpetuate among those who are vaccinated, I'm wondering if it might not actually be better than nothing.

Today's news is not about a vaccine that will be distributed, it's about having enough success that it may suggest avenues for further study. I think everyone involved recognizes this as a qualified success, but in HIV vaccine research we haven't really yet had even a qualified success yet, so this is very exciting.

Like the lack of infection may have been more to do with safe sex practices, etc., than the vaccine itself. I would think that a working vaccine would have a much higher rate of prevention.

Many people are puzzled by this, and by the mechanisms by that are in operation here.
posted by OmieWise at 9:13 AM on September 24, 2009 [1 favorite]

More specific numbers from the NY Times:

Col. Jerome H. Kim, a physician who is manager of the army’s H.I.V. vaccine program, said half the 16,402 volunteers were given six doses of two vaccines in 2006 and half were given placebos. They then got regular tests for the AIDS virus for three years. Of those who got placebos, 74 became infected, while only 51 of those who got the vaccines did.

Although the difference was small, Dr. Kim said it was statistically significant and meant the vaccine was 31.2 percent effective.

posted by smackfu at 9:14 AM on September 24, 2009

AIDSVAX is, at a handwavy level, a gp120-targeting vaccine, so it would be widely effective, rather than strain-specific. It looks like the ALVAC-HIV vaccine (vcp1452) has genes for the HIV core as well as its sugary coating. I thought this was it, but it appears to be targeting the wrong part of the world.

"2) how can you test for a lowered risk of infection without attempting to infect someone?" — you give some people placebos, some the real deal, and then you let them do what they normally do, which is go out and get laid, then compare the rates of infection.

I really hope they get somewhere with this.
posted by adipocere at 9:16 AM on September 24, 2009

1) why the need to clarify that the volunteers were heterosexual?

Unlike prior vaccine trials, this one used volunteers from the general population rather than at-risk groups like homosexual men, prostitutes, and injected drug users.
posted by jedicus at 9:19 AM on September 24, 2009

I anticipate that conservative groups would oppose this vaccine in much the same way and for the same reasons they oppose Cervical cancer vaccines.
posted by jefficator at 9:22 AM on September 24, 2009

Ah, thanks smackfu, adipocere, and jedicus - it makes more sense now.
posted by filthy light thief at 9:30 AM on September 24, 2009

They way I understand the trial is that it works like this: You take a large group of people, split them into two groups (that are sufficiently similar to each other, same number of drug users, same socioeconomic backgrounds, etc). Administer the vaccine to one group and a placebo to the other. After a sufficient amount of time, compare HIV infection rates in the two groups. They should be similar minus the effectiveness of the vaccine. If you have a statistically significant difference, then you have a possible vaccine.
posted by borkencode at 9:30 AM on September 24, 2009

There are lies, damn lies, and statistics. So because 20 fewer became infected they give credit to the vaccine? What if those 20 were too busy playing WOW to have sex? What if those 20 had only one partner while the rest had multiple partners? What if those 51 and 74 went to see prostitutes and the remainder did not, what if those 51 and 74 lied, and were actually gay men with a greater chance of getting the disease? I think it too soon to give too much credence to this study.
posted by Gungho at 9:41 AM on September 24, 2009

Gungho, one imagines that this study represents some statistical significance in relation to other studies that used similar populations. On a macro scale, the odds of individuals being able to thwart study results by lying and being exceptionally more promiscuous should be pretty low.
posted by Think_Long at 9:45 AM on September 24, 2009

Gosh I wonder if they thought of those things before conducting this years-long expensive study
posted by showbiz_liz at 9:46 AM on September 24, 2009 [8 favorites]

Gungho: This was a blind study. The participants didn't know whether they were in the control group or the experimental group. The odds that your hypothetical set of liars or miscreants all ended up in one group or the other so as to affect the outcome of the study are extremely low. In this case p < 0.04 and a 95% confidence interval.
posted by jedicus at 9:53 AM on September 24, 2009 [1 favorite]

Gosh I wonder if they thought of those things before conducting this years-long expensive study

DID YOU KNOW THAT ALL CLIMATOLOGY IS DISCREDITED BECAUSE WEATHER RECORDING STATIONS ARE IN CITIES NOW?
posted by Pope Guilty at 9:55 AM on September 24, 2009

On a macro scale, the odds of individuals being able to thwart study results by lying and being exceptionally more promiscuous should be pretty low.

As low as 20 out of 2000???
posted by Gungho at 10:02 AM on September 24, 2009

Gungho: There are lies, damn lies, and statistics. So because 20 fewer became infected they give credit to the vaccine? What if those 20 were too busy playing WOW to have sex? What if those 20 had only one partner while the rest had multiple partners? What if those 51 and 74 went to see prostitutes and the remainder did not, what if those 51 and 74 lied, and were actually gay men with a greater chance of getting the disease? I think it too soon to give too much credence to this study.

The whole point of doing a study with a relatively large sample size and random assignment to control and sample groups is that these extraneous variables should wash out in the background noise.

kittens for breakfast I'm not a scientist or anything here, but 32% just sounds...odd. Like the lack of infection may have been more to do with safe sex practices, etc., than the vaccine itself. I would think that a working vaccine would have a much higher rate of prevention. I would be happy to have come to the wrong conclusion, obviously.

Certainly, which is why this is not being presented as a "working vaccine." It is the first major breakthrough in the attempt to design a vaccine for HIV in a long time, which makes it a very, very, very big deal. While we are still years away from a working vaccine, we now have promising direction beyond just engineering antigens onto innocuous viruses and hope they trigger a strong enough immune response.
posted by KirkJobSluder at 10:05 AM on September 24, 2009 [1 favorite]

Local physician: HIV/AIDS cure getting little publicity
posted by hortense at 10:09 AM on September 24, 2009

As low as 20 out of 2000???

Well, higher actually (20:2000 is 1%). There was a 4% chance that the result they got could have been because of chance alone. That's pretty unlikely, but not as unlikely as one would like before embarking on a massive vaccination program. But because the measured effect was somewhat weak, more studies will necessarily be done as the vaccine is refined to be more effective. If the effect was due to chance alone, those followup studies should reveal that.

In short: the researchers know what they're on about and have not overstated their findings, even if some of the reporting has been a little breathless. If you continue to have trouble believing this, learn statistics.
posted by jedicus at 10:11 AM on September 24, 2009 [5 favorites]

Local physician: HIV/AIDS cure getting little publicity

That case study got quite a bit of publicity, actually. Not much has come of it because, for starters, finding bone marrow donor matches is very difficult, and the problem is dramatically compounded if the donor also has to have the very rare mutation that gives immunity to HIV. It's more of a proof of concept that gene therapy could indeed work to treat HIV.
posted by jedicus at 10:19 AM on September 24, 2009

So because 20 fewer became infected they give credit to the vaccine?

I'm not a statistician, but I am a programmer so I tried out a quick simulation to see what kinds of swings would be expected with this sample size.

I assumed that the infection rate in the control group is accurate (around 74 per 8201 people), and did a simulation of 10,000 runs, giving everyone in both groups a randomized 74 in 8201 chance of getting the disease. This would simulate what would happen if the vaccine was completely ineffective (i.e. if both groups had the same infection rates). Of those 10,000 simulated runs, in only 310 cases (around 3%) the control group had a 23 person or greater reduction in infected cases.

So although the 32% reduction rate might not be completely accurate, the chances of the reduction being due to random differences in the sample groups is very low. Systematic failures in the testing process would probably be a more likely cause for the results to be inaccurate, which is why part of the process of scientific studies is to have other people attempt to reproduce the results.
posted by burnmp3s at 10:35 AM on September 24, 2009 [11 favorites]

So, anyone who understands epidemiology or whatever relevant field tracks the spread of disease want to explain how this vaccine, if administered to the whole population, would reduce the spread of the disease, assuming everything else stays the same (meaning people don't get a sense of immunity and practice less safe sex)? If anyone feels like crunching two sets of data, I'd like to hear how it would work for an industrialized nation and a developing nation.

I get the feeling that 32% of the people being immune would have a ripple effect because of there being less vectors, but I don't know if it'd be a big decrease in cases or a small one.
posted by mccarty.tim at 10:55 AM on September 24, 2009

assuming everything else stays the same (meaning people don't get a sense of immunity and practice less safe sex)

I think that would be a major flaw in your model right there, at least if you wanted to be able to use it as the basis or rationale for public policy. Once a vaccine is found that's determined to be effective via blind studies, then limited trials will have to begin where the participants know they've gotten a vaccine, to see what the net effect is, inclusive of changes in behavior.

I think that will be an interesting and important study to perform, because I suspect there will be a lot of opposition (at least in the US) from social-authoritarian groups which will argue against vaccine distribution on the grounds that it does more harm than good because of behavioral changes. I would anticipate the arguments being similar to those against condom distribution. Of course, those arguments will be nothing but a fig leaf for the real opposition* to an AIDS vaccine, but they need to be put down good and hard just the same.

* The real reason being, of course, that AIDS disproportionately infects homosexuals, drug users, and minorities, making it practically an instrument of divine will in the minds of some people I've had the misfortune of being in the same room with. In the same way that authoritarians claim many reasons for opposing birth control and contraception, but in the end have actions entirely consistent with slut-punishing via pregnancy, I fully expect there to be many stated reasons for opposing an AIDS vaccine but for the actual actions of those opposed to be entirely consistent with a desire for racist/homophobic genocide.
posted by Kadin2048 at 11:28 AM on September 24, 2009

burnmp3s: Just to point out, what you did is called a Monte Carlo or resampling test. Classical statistics does the same thing using calculus to predict the behavior of an infinite number of ideal simulations. Classical statistics tends to fall on the assumption that its theoretical simulators result in ideal curves, which may or may not be true of the phenomena you want to study, but is usually "good enough."

mccarty.tim: Well, I think it's pretty clear that the results returned come nowhere close to the level of efficacy that justifies widespread implementation. These results are in no way practical, and no one is going to ethically or politically greenlight mass production of this particular vaccine based on these results.

What makes this groundbreaking news is that earlier attempts to produce a vaccine have had disappointing human trials. One vaccine study was aborted because it may have increased infection rates, and another vaccine study was aborted because preliminary results showed no effect. It was reasonable to consider that we might never have a vaccine, that the complexities of HIV and its interaction with the human immune system might serve to make a prophylactic vaccine impossible to develop.

A weak positive result based on the interaction between two different types of vaccine is a significant step forward. Researchers can now go back to the animal and in vitro models and puzzle out exactly what's going on that makes the interaction more effective, and hopefully produce a better prophylactic vaccine that could be ethically used across an entire population.

This is rather like the early penicillin therapies. The early trials of penicillin were not huge successes, but they were promising enough to justify further research and experimentation. Perhaps this will be another frustrating dead end. But it's a promising development.
posted by KirkJobSluder at 11:34 AM on September 24, 2009

* The real reason being, of course, that AIDS disproportionately infects homosexuals, drug users, and minorities, making it practically an instrument of divine will in the minds of some people I've had the misfortune of being in the same room with. In the same way that authoritarians claim many reasons for opposing birth control and contraception, but in the end have actions entirely consistent with slut-punishing via pregnancy, I fully expect there to be many stated reasons for opposing an AIDS vaccine but for the actual actions of those opposed to be entirely consistent with a desire for racist/homophobic genocide.

I really feel the need to categorically disagree with this interpretation on two grounds.

The first is perhaps merely semantics: AIDS does not disproportionately infect any group. That wording implies that AIDS discriminates in its own selection of victims. This is false. A virus does not decide to infect its host based on such constructs as sexual orientation. Your point is taken, however. There are groups that suffer from AIDS disproportionate to their numbers within the general population.

The second point is made not simply to be contrarian, but by way of refuting the notion that racism and homophobia alone are responsible for a lack of concern with AIDS prevention in some communities. This implies an irrational behavior that cannot be reasoned with or debated. On the contrary, many who hold the view you lament would argue that disease is a by-product of certain aspects of "lifestyle." I have heard extremist consider AIDS to be divine retribution, yes, and so I won't argue that point. But I believe far more people whom you portraying monolithically would argue that disease comes as the logical conclusion of failing to protect yourself by refraining from unprotected sex, drug use, and other such dangerous activities. Now if they are simply irrational hater-mongers, nothing can be said to them. But there is, I think, space for amendment of thought when people are reminded of rape and informed of such outrages as the God-only-knows-how-it-started-but-common-in-Subsaharan-Africa notion that sleeping with a virgin will cure AIDS.

I believe if we give people a little more credit and a lot more information, they will make better decisions. That involves not only those prone to AIDS in particular communities, but also those prone to harmful opinions about AIDS in other communities.

Let's spread the understanding a little more liberally.
posted by jefficator at 11:48 AM on September 24, 2009

Running their results through a chi-square test, it looks like their confidence interval is just barely over 95% (which is usually the threshold for calling something a "success" in the drug world.)

Another way to look at it: In any 20 such studies, there will tend to be one "success", even if all the studies were accidentally using sugar pills for both the control and experiment groups.
posted by Teppy at 11:53 AM on September 24, 2009 [1 favorite]

The first is perhaps merely semantics:

I think it was definitely semantics. Maybe this would have been a bit better: AIDS disproportionately AFFECTS homosexuals, drug users, and minorities
posted by Think_Long at 11:55 AM on September 24, 2009

burnmp3s: Just to point out, what you did is called a Monte Carlo or resampling test. Classical statistics does the same thing using calculus to predict the behavior of an infinite number of ideal simulations.

Yeah I know. Nobody had posted a statistical calculation how reliable those results were given that sample size, so I just wrote a a few lines of code to get an approximation. Maybe some day I'll bother to learn how to do it the real way (I never took statistics in school, although I did take courses on probability and combinatorics).
posted by burnmp3s at 11:56 AM on September 24, 2009

By far the most common transmission of HIV is through heterosexual sex. My guess is that the researchers just chose to select a specific pool, heterosexuals, to narrow down the uncontrolled variables. On top of that, in Thailand, you probably have trouble finding a large enough population willing to identify themselves as gay. It would probably be much harder to create a vaccine that worked against HIV derived through IV drug use. You can inject live infected cells, not just naked virus for the immune system to bring down.
HIV is not that easy to transmit in a given sex act. A 31% reduction in the transmission success might be good enough to stop the growing tide of the epidemic (which unfortunately occurs when more infected people die off than new people get infected).
posted by dances_with_sneetches at 12:56 PM on September 24, 2009

By far the most common transmission of HIV is through heterosexual sex.

Depends on where you're talking about. In Africa that's probably true. In the United States it isn't close to true.
posted by Justinian at 1:12 PM on September 24, 2009

(or at least it wasn't true historically speaking. Transmission rates have been changing over the last couple years)
posted by Justinian at 1:13 PM on September 24, 2009

Probably depends on whether "common" is referring to raw counts or to rates.
posted by smackfu at 1:25 PM on September 24, 2009

Well, it's indisputable that transmission rates are much lower for heterosexual sex. But heterosexual sex wasn't the most common transmission of HIV in the United States even in terms of raw count. My guess is that it still isn't, but I wouldn't swear to it since I haven't seen the numbers recently.
posted by Justinian at 2:14 PM on September 24, 2009

Here are some recent statistics on transmission by risk group for the US. One third are heterosexual by this count. The take home lesson for the Moral Majority types is: promote lesbian sex.
As for saying above, "By far the most common transmission of HIV is through heterosexual sex" of course I wasn't referring to the US. The US is a small part of the AIDS epidemic.
posted by dances_with_sneetches at 2:57 PM on September 24, 2009

Thanks for the recent stats, dances, that's very helpful.
posted by Justinian at 3:19 PM on September 24, 2009

I should have said something like "rates of infection are disproportionately higher among..." rather than "the virus infects...". I didn't mean to imply that the virus itself somehow acts differently once in the body of, say, a homosexual versus a heterosexual person. Only that because of the methods of transmission, exposure and thus infection rates are higher in certain groups—and those groups tend to be the ones that are frowned upon or seen as "immoral" by certain hateful, holier-than-thou sections of the population.

I hope, for all our sakes, that it's only a small core of hatemongers who really do believe the AIDS-as-divine-retribution thing…I guess we'll see if and when a vaccine is developed. It can't come too quickly.
posted by Kadin2048 at 12:04 AM on September 25, 2009

They revealed their full data today and it's not looking quite as rosy. From the Guardian:

• The vaccine did not protect those at high risk of HIV infection, such as sex workers and intravenous drug users
• The protective effect was greatest in the first 12 months and then seemed to diminish
• When those who did not get all six vaccine shots were taken out of the analysis, the positive result was statistically insignificant

The third point seems to be the kicker:

But many volunteers did not get all six vaccinations, taking the numbers down from around 8,000 in each group to around 6,000. Among those people, there were 50 infections on placebo and 36 on the vaccine, which gives an efficacy of 26% but is not statistically significant (meaning it could happen by chance).

posted by Rhomboid at 11:05 AM on October 20, 2009 [1 favorite]