HIV superinfection in chronically-infected patients reported

[keyite]

Aidsmap brings this story.

It was previously widely believed superinfection occurred only within three years of initial HIV infection - this study finds evidence of superinfection occurring more than five years after patients first became infected with HIV.

[newt]

This is methodologically and descriptively flawed and statistically underpowered, and proves not one case of superinfection that does not have an equally reasonable alternative explanation (change in initial viral population which may comprise more than 1 strain).

- matt

[bobino]

Matt,

Would you mind expounding on your comments for the statistically challenged among us?  I understand that the study examined only 36 individuals, of whom only seven showed evidence of reinfection, so it's a very small study.  Other than that, though, what do you find flawed in it?  I don't understand the parenthetical comment that closes your post.

Sorry if I'm being obtuse, but I'd really like to understand.

Thanks,

John

[newt]

With pleasure..

Paper in question is online in full here

First, let's start with the paper title

Chronic HIV-1 Infection Frequently Fails to Protect against Superinfection

Ahem, Chronic HIV-1 infection may not protect against superinfection: more research needed would be a fairer conclusion.  I am always alerted to strong conclusions based on small samples with no control.  In this case the investigators are looking for reinfection after 1 year, so you have to consider selection bias and the like (cos you tend to get what you measure...or want to measure), which leads me to...

References...

For one, they don't reference any of Julia Marcus' work on reinfection.  Something of a literature bias.

Higher Frequency of Apparent HIV-1 Superinfection -- Sequentially Expressed Dual Infection (SEDI) -- in Recent Infection Compared to Chronic (PDF poster)
Robert M Grant, J Jeff McConnell, Julia L Marcus, Christophe Kreis, Gerald Spotts, Teri Liegler, Rosaleen Brennan, and Frederick M Hecht, Gladstone Institute of Virology and Immunology/University of California, San Francisco

This paper contains "Leon", where a second virus appeared shortly before a collapse in CD4s.

"One case described here would have met criteria that other groups use for "superinfection" in that a virus population at baseline was replaced by a second highly divergent viral variant. However, analysis of the screening sample indicated that the subsequent virus was present at baseline. This case demonstrates how the dynamics of dual infection can lead to the appearance of superinfection."

Marcus' work suggests a null hypothesis, namely reinfection is not really reinfection but a change in the initial viral population (for some reason).  To make strong statements, indeed anything more than suggestive, about reinfection you need to disprove the null hypothesis.  This paper doesn't even mention it. Perspective bias. Bad methodology.

The sample size is 36.  Underpowered.

The sample size is 36.  Phylogenetic analysis is not robust at small numbers in terms of viral history. (It's good at a population/epidemiological level).  Phylogenetic analysis would not be admitted in a British court of law as robust evidence showing a chain of infection, for example, between two individuals, even if their genotype was <10% different.

The assumption that the 7 individuals were not coinfection with 2 (or more) strains of HIV is justified via unpublished data.  Therefore cannot be examined by the reader.  Bad methodology.

Maximum-Likelihood Phylogenetic Trees is an interesting an statistically robust mathematical modelling process at a population level but at an individual (single case) level has not been validated.  Indeed, it was not designed for this purpose. Bad methodology. Furthermore, it's an advanced technique and a comparative analysis by another method would be sound, to cross-reference the results << bit of boys and toys going on

To confirm an infection you need to examine both recipient and source.  No source testing is included in the study (well, it would be hard with working girls...). Therefore you simply cannot say that the interesting data on genetic origin/change is down to reinfection. Its bad, bad, leap of faith science. Bad methodology << Of course this does not mean they are not cases of reinfection, just unproven either way.

So, to me, interesting data, but case unproven, and written up with a large degree of naivety on the researchers' part.

Plus, the real question is, so what if reinfection does happen, when, if ever, is it clinically significant (protective, destructive -- it could be both, who knows really)?

Hope this clarifies why I said what I said.

- matt 

[NewYorkKat]

Would superinfection only happen if the HIV positive individual would engage in unprotected sex with another HIV positive person?

If so, the problem would stem from more unprotected sex.

[bobino]

Matt,

Thanks so much for your explanation.  Your knowledge of HIV is clearly far superior to mine, so much so that I had to read your second post a few times before understanding it. 

I had also wondered about the problem that you put so succinctly -- how does one distinguish reinfection from what I'll call "simultaneous infection"?  That is, isn't it entirely plausible that someone might be exposed to more than one strain of HIV upon initial infection?  I think I recall reading somewhere that one can have different strains of HIV in the body, but one may be "dominant" for a certain period, but the other one can later emerge from a dormant state.  I guess what is needed is a way to determine when a particular viral strain was acquired, and it would seem that that would require some kind of partner tracing.

So thanks again for the explanation.  This issue is important for those of us living in San Francisco, where so many gay men are HIV+.  Just another reason to practice safe sex.

John


(To NYKat:  I think you're right.  Superinfection is only possible if an HIV+ individual engages in unprotected sex with another HIV+ person.  So, yes, the problem stems from unprotected sex (or sharing needles).)

[krh]

It also seems that superinfection occurs only with those who are not on therapy.  For those on fully surpressed therapy, my Dr. had told me that superinfection had never been recorded.  Again, he said "recorded" and said it is possible but not probable.

[bimazek]

if i had a dollar for every time an HIV professional tried to put the fear of god into me about super infection i would be rich, it is like they use it as a hammer to try to stop people from having any sex, and the fear and worry and anxiety esp. when newly diagnosis ed is ridiculous to add that burden onto people

then i talk to all my hiv friends and they say the opposite that it is hard to catch again and they have been having bb sex with other poz guys for years

it is very confusing upsetting

anyway

there are tons of other std's

so be careful

but  ....  this study raises tons of questions

like

was the new virus just a mutation of an old one?

is the person catching a virus that is now resistant to meds and before they did not have any resistance to meds

this is important

[NewYorkKat]

Superinfection exists. Remember, that viruses and bacteria are living things and like all living things, they can adapt and change.

being poz and putting yourself in more danger by having more unprotected sex doesn't;t solve or make things better. It makes things a lot more worse.

They are POZ people with more than one infection such as Hep C or herpes for instance.

Your friends. sorry to say, are not sure themselves and if they were correct, HIV would be more easier to manage than previously thought.

Using protection is the only way of life. Condoms have to be worn, like or not. If we did in the first place, think you or me or anyone else would be on this board?

Getting a positive result is scary but hearing you got more than one infection on top of your HIV can be more scary.

Just use precautions and use a condom!

[bimazek]

I use condoms but

i would like to hear of ten people on this board who had the Phenome and Genomes and they had wild type virus and then they had normal sex without a condom perhaps because they were meth users or something

then they found that they had resistant virus later on

would it be fair to say that 99% of the resistance and death caused by resistance to meds is caused by the virus evolving in response to the meds?

and now we have a treatment strategy that is cause more mutations so virus collapses  KP-1212

[Miss Philicia]

i would like to hear of ten people on this board who had the Phenome and Genomes and they had wild type virus and then they had normal sex without a condom perhaps because they were meth users or something

Can't wait for replies to roll in on that request.

[newt]

Superinfection does exist

..and the moon exists, but most of us will never go there...

- matt

[dixieman]

Why worry about reports... take care of yourself... somethings gonna get you one day think in the now...

[HIVworker]

In addition to the points Matt makes, I worry about their ability to detect viruses at baseline.

"We calculated that by using ssPCR we had a >99% probability of detecting a strain present at 10% prevalence, a 92% probability of detecting a strain present at 5% prevalence, and a 40% probability of detecting a strain present at 1% prevalence."

While this sort of PCR analysis is interesting at these levels of detectability, it sensitive enough to suggest that the virus was not present at baseline - even if they get a negative result for a virus that appears later on. It also depends on the VL of the patient when the sample was taken - although in all honesty the VL isn't low enough to suggest PCR bias.

While the detection levels seem impressive, in someone with a VL of 100,000 copies/ml, 1% of the virus is still a whopping 1000 copies/ml. If you consider the body has about 5L of blood then 1% of the virus in the blood of an infected person with a VL of 100,000 is a 'not so tiny' 5,000,000 copies per body...and given that the whole body reservoir consists of not just the blood compartment this is likely an underestimate.

So I don't really quibble with their phylogenetic trees, I do quibble with their ability to detect the virus at baseline. There is no way that they can say the virus wasn't there.

Superinfection is a serious issue and demands exploration. This paper does a better job than most - although it is still flawed.

Why is superinfection important? Well, people often talk about recombination of viruses to make a fitter strain. That could happen. However, superinfection with a drug-resistant strain would be more of an issue. If you had someone with a known drug-resistant virus having unprotected sex with someone with HIV who was on meds and didn't have that drug-resistance and suddenly developed a drug-resistant virus that didn't look like their own but more like their partners then that would be more convincing.

R

[bimazek]

hivworker you have great post keep posting very informative

i am confused because it seems to me that haart is what causes the resistant strains, anyway, i had as safe sex as i could for 25 years and then in the 26th year of this epidemic i made a mistake, no that i have all the weight and pain and emotional upset about being hiv, i am trying to still have safe sex,

one question

would it matter if a person was on HAART or not for that person to get a resistant strain?

in other words wouldnt HAART help select FOR a resistant strain if the person was on HAART and had unsafe sex?

or  would HAART help to keep it from catching even if it was resistant

[HIVworker]

If they were on HAART and it was active (low viral load in them) then there would be few (if any) virions carrying resistance to the HAART they were on. If their HAART was failing it would be more likely. That doesn't mean that they wouldn't have other virions carrying resistance to drugs other than those in their HAART.

If they were NOT on HAART they could still generate a mutation that had resistance to your HAART -  but the combined likelyhood of them doing that AND you receiving it would seem very slim.

R

[megasept]

Thanks for the question and Matt's (typically) detailed clear logical response.

I worry about "super-infection" as much as Anthrax in my mail. Much ado about nothing.

  -megasept