[MechEStudent]

HIV evolves with amazing (super-villian-ic) speed. Many "cures" attack 99% of cells, or similar. Resistance is evolved during treatment, and is usually uniform within just a few months. It mutates/evolves so fast that given just virus samples, it is possible to determine sequence of infection.

I think it needs to kill more than 6-9's before it has a good chance at being a cure.

[Iv]

I don't think this is relevant in that case. I am not an epidemiologist but I worked with people who were studying HIV. My understanding is that while the HIV mutates fast, it does not create new strains during the treatments.

A given strain can have some variability but will share general traits.

HIV strains are named and numbered: https://en.wikipedia.org/wiki/Subtypes_of_HIV

An efficient treatment against a specific strain would save millions of lives. An efficient treatment against the 99% of them that are the most prevalent would basically solve the epidemics.

The problem is that people often end up with more than one strain. This is why doctors tell HIV+ people to continue protect themselves and were worried about the appearance of "free-sex parties" for HIV+ individuals. Most people do not differentiate between strains but if you got HIV from a partner, you may have had only a part of their strains, you should not have unprotected sex if you want to maximize your chances.

[api_or_ipa]

Reminds me of a lesson about misleading statistics I tell my nephews. We've all see the hand sanitizers that advertize killing 99.99% of germs, but what does that actually mean? 99.99% equates to 1/10,000 survivors and assuming for each generation the population doubles means it takes just 14 generations for the original colony to be back at full strength again. For extra creepy effect, e coli can reproduce in optimal conditions as fast as every 20 minutes which buys you only a couple hours of hygiene[0].

Given this drug kills 99/100 HIV virus units, then this really only helps you for log2(100) or ~7 generations

The math is pretty easy, but the results are something that always surprise me: it turns out humans are pretty bad at estimating exponential growth.

0: real life conditions are probably slower than this, I concede.

[apta]

> which buys you only a couple hours of hygiene

Which is long enough for you to complete your lunch! :-)

[Latty]

Is that even the claim they are making? I honestly never thought about it, but the claim could easily be "we kill 99.99% of germ strains" or "we kill 99.99% of germ cells".

[jamesrcole]

I think the problem is that people tend to focus on the size of the numeric probability, as if it was intrinsically meaningful. And don't think to consider contextual details like in the example you give. It's the combination of the probabilities to the contextual details they're applied to that really matter.

Another illustration of that is, if 99.99% equates to 1/10000 survivors, the meaning of it depends on how many bacteria exist on the area in the first place. It has quite a different meaning if the typical number is 10000 vs if it's 100000000

[ttflee]

> e coli can reproduce in optimal conditions

Then avoid dipping your hand into beef stew for too long. Most of time, germs grow like sigmoid function.

[smitherfield]

The idea is that it kills 99% of them, and then your immune system is up to the task of killing the remaining 1%.

[DrScump]

But it's not a simple, linear battle.

That most virulent 1% then multiplies and infects systemwide... and adds new mutations all the time.

[smitherfield]

It's not the most virulent 1%, it's 1% which have unusual mutations that cause them to be resistant to this treatment, but which usually reduce their fitness in other respects. (This can be seen in humans as well, e.g. malaria resistance->sickle-cell). The fastest-multiplying viruses are also (obviously) going to be the most common, in the absence of treatment.

[jondubois]

This is really fascinating how being really good at something often correlates with being less good at other things (in this case antibiotic resistance seems to be inversely proportional to the potency of the bacteria).

This is true not just for bacteria but also for other living creatures; for example animals that have lots of children have very short average lifespans.

It makes me think that every living species must be at the absolute cutting edge of its capabilities genetically-speaking.

[46886532558]

Natural selection tends to work on traits that are relevant till the ability to reproduce ends. If the trait can dominate by rapid reproduction, life span does not matter and robustness is not selected for. Generally it is not an inherent trade-off between certain traits, apart from energy consumption.

[xelxebar]

My physics brain is now imagining some conserved quantity like the fitness density of a bacteria population. Lol

[therein]

Interesting point. I hadn't considered unusual but useful mutations they accrued might come with reduced fitness in other regards.

[acjohnson55]

It's like having superpowers, but with physical abnormalities. Like so many things, this was explored in the comics!

[nebabyte]

the marketing idea?

[autokad]

hand sanitizers aren't to replace soap, where as soap kills and washes surviving germs away

[delbel]

I'm not familiar with any details about this vaccine or other therapies, but from what I understand is when you have a partial solution or near total solution like this, you combined it with another therapy so you have two methods of action to dealt a "kill shot" to the target organism or virus. This combined therapy strategy is used in killing Borrelia type infections. Basically they combined two different antibiotics at the same time. Alone they are effective at relapse rate of say only 5% after 14 weeks, combined they can get near 1% or lower. Also when an organism evolves, its genetics can drift causing mutations that are disadvantageous. There a chance the mutation could make it more vulnerable to something just as well as immune to something else, and also there has to be consideration that there is a scenario where that can happen at the same time.

[cbhl]

The article says they already combined three different antibodies to get the 99% rate.

[pasbesoin]

HIV was/is an early hint at the computational complexity we face in understanding biology.

Not that we should consider in insurmountable nor ungrasp-able. But that we -- or parts of our society and professional institutions -- should lose some of our arrogance.

On the upside, that complexity also harbors enormous opportunity. Just, don't be simplistic about it an let your ignorance take you into disaster.

[phyllostachys]

I'm always amazed at how many doctors I've encountered behave as if their knowledge is the extent of medical practice. I get it though, research is hard to stay on top of and isn't linear to results.

Granted, also, that there are lots of doctors who do follow research.

[aswanson]

Most doctors are decision-tree driven, following the rules. If you get a truly insightful one, recommend the f out of him/her.

[jk2323]

While your argument has some validity, the statement said "New antibody attacks 99% of HIV strains" not "New antibody attacks 99% of HI Viri (of a given strain)". The mutation would have to occur faster than the virus dies off.

[wickawic]

I'm sure that it is much more complicated than this, but I wonder how long the virus can remain immune to all previous '99%' cures? Maybe eventually we could come up with enough of these and then administer them all at once.

[chimeracoder]

> I'm sure that it is much more complicated than this, but I wonder how long the virus can remain immune to all previous '99%' cures? Maybe eventually we could come up with enough of these and then administer them all at once.

That's basically what HAART (the technique we've used to treat HIV successfully for the last 25 years) is. People take three antiretrovirals simultaneously. They're designed such that, for the virus to adapt to one, it has to make itself more susceptible to one of the other two.

[hn_throwaway_99]

> They're designed such that, for the virus to adapt to one, it has to make itself more susceptible to one of the other two.

That's not really accurate. HAART works because the 3 drugs are taken simultaneously, and it's extremely unlikely for a virus to get random mutations that confer resistance to all three in one generation.

[chimeracoder]

> That's not really accurate. HAART works because the 3 drugs are taken simultaneously, and it's extremely unlikely for a virus to get random mutations that confer resistance to all three in one generation.

That is literally exactly what I said. The drugs are taken simultaneously, and it is very difficult to develop resistance to all three simultaneously, because becoming more resistant to one means becoming more susceptible to another.

[hn_throwaway_99]

Not sure if it's just miscommunication, or you misunderstand how HAART works, because you said "That is literally exactly what I said", and then in your next sentence repeat the statement that is false, "becoming more resistant to one means becoming more susceptible to another."

For example, suppose you take ATripla, which is a combination of efavirenz, emtricitabine, and tenofovir. If a viral particle has a random mutation that confers resistance to, say, efavirenz, that mutation does not make the virus more susceptible to emtricabine or tenofovir. The fact is the virus was already susceptible to those two drugs (that's why you take tests that specify your viral strain's pre-existing resistances to determine the best drugs for you before you start HAART), so it still is unable to replicate.

It's just a matter of statistics, not that resistance-granting mutations naturally make the virus more susceptible to other drugs.

[carlisle_]

I believe that's actually somewhat how AIDs treatments have worked. I'm not sure if all are like that but it's pretty much the exact logic.

[esm]

Yep. Triple therapy is now standard in HIV treatment. The odds of developing mutations to 2-3 agents simultaneously are very low

[skygazer]

When I was a teenager, I wrote a connect four game, and my brother would play it. Every time he'd win, I'd jump back into the code. I realized it was a fruitless cat and mouse (not only because the game is trivial) but because of the pacing of my "fixes" -- if I could have had the final version completed before he first played, he might have been stymied, but the iterations seemed to give him an easy path -- he only needed to solve one problem at a time.

[bjterry]

Why don't we deliver all antibiotics as a triple? Wouldn't that eliminate antiobiotic resistance?

[DrScump]

Antibiotics kill bacteria, not viruses. And that includes desired strains.

[c3534l]

Probably for longer than we can keep coming up with chemicals that selectively kill it. After all, if we're not killed by the chemical, then that means it's possible to survive it.

[fghvbnvbnfe]

That's not really a meaningful argument. You're talking about a large multi-cellular organism that has tissues, organs, and all that and comparing it to something that doesn't even have cells.

[patryn20]

This is the fundamental problem with treatment of rapidly evolving diseases. The hope is that if you get a "good enough" treatment and get it out to the population rapidly enough, the disease won't have a chance to evolve and spread further. Herd immunity will bring it under control.

I think the problem with a widespread, highly studied virus like HIV is that a "good enough" treatment isn't "good enough." Any treatment will be deployed in a limited manner to test the efficacy, which will in turn give it a chance to develop a resistance and spread.

That having been said, if the treatment CAN treat 99% of current infections, it will be a boon for at least a couple generations of gay and minority communities in the developed world.