I Vant To Drink Your Blood!

August 18, 2008 4:26 PM

RidiculousHypotheticalQuestion: If someone were to drink a mouthful of HIV/AIDS infected blood, what would the chances of becoming infected be? (This was spawned from one of those ridiculous back-and-forths, about infected candy, at work.)
posted by P.o.B. to health & fitness (12 comments total)

It would depend on whether the person drinking the blood had any cuts in their mouth. But seriously, did you even try googling for this?
posted by nerdcore at 4:33 PM on August 18, 2008

The rates for exposure from various kinds of fluid exchange range from .005% to 90%. Eliminating blood transfusions and childbirth as outliers gives a range of .005% to .67%.

Assuming that the drinker has no cuts, I would guess it's (at worst) equivalent to multiple instances of intercourse, as the exposure is all along mucous membranes. Taking the highest transmission rate for intercourse (.67%) and multiplying by the length of the GI tract (~393 inches) divided by the (ahem) length usually exposed during intercourse (~6 inches) gives 44%, which is pretty high.

I'm not a doctor or a biologist, though. I would suggest asking an HIV researcher.
posted by jedicus at 4:56 PM on August 18, 2008

I have no expertise in this matter but even I can say that the 44% chance that jedicus pulled out of his butt is...unreliable...to say the least. First, that's not the way you calculate the probability of a compound event. Second of all, the GI tract is far from a homogenous environment throughout its length -- who knows which sections are more or less vulnerable to viral infection?
posted by randomstriker at 5:08 PM on August 18, 2008

Yeah - google.

Really, because I don't see a definitive answer on this anywhere by just googling it. If somone knows specifically that would be appreciative, rather than me wading through 20 different web pages of AIDS on fountain drinking myths.
posted by P.o.B. at 5:13 PM on August 18, 2008 [1 favorite]

Taking the highest transmission rate for intercourse (.67%) and multiplying by the length of the GI tract (~393 inches) divided by the (ahem) length usually exposed during intercourse (~6 inches) gives 44%, which is pretty high.

This math is bogus. Using your model, if your GI tract was 1200 inches long, you'd have a 134% chance of getting infected. It also seems unlikely that anything other than the length of the esophagus counts, given the pH of stomach acid.
posted by 0xFCAF at 5:26 PM on August 18, 2008 [1 favorite]

You'd also have to account for the person possibly throwing up immediately, which might cause micro-tears along the esophagus and dramatically increase risk.
posted by Benjy at 5:37 PM on August 18, 2008

There isn't an answer to this. Establishing rates of infection for any transmission vector requires large epidemiological studies. So the rate of infection from, for example, vaginal intercourse, is established by looking at the actual rates of infection in people whose exposure is likely to be through vaginal intercourse. It's not calculated from first principles.

Drinking blood, while it does happen, is nowhere near common enough for these kinds of studies to be worthwhile. However, the general consensus is that rates of HIV transmission through receptive oral sex are very low. Obviously a mouthful of blood is a larger volume of liquid than the average ejaculation so the transmission rate for drinking a mouthful of blood would presumably be slightly higher than for receptive oral sex, but not as high as other sexual activities involving fluid exchange.
posted by xchmp at 6:43 PM on August 18, 2008 [1 favorite]

My estimate was meant to be an upper bound, not an exact answer, which is why I ignored the effects of the stomach (as well as the possibility of cuts).

0xFCAF and randomstriker are right about the probability calculation. The events are non-exclusive and independent, so they do add, but I forgot to subtract the probability of the intersections. Doing that I get 30%, but someone with more statistics should check that.

Again, I was trying to establish a loose upper bound. The exact answer is, as xchmp pointed out, essentially unknowable, so a guess at upper and lower bounds is as good as it gets.
posted by jedicus at 6:58 PM on August 18, 2008

Not precisely the same scenario, but the CDC says: Even if small amounts of HIV-infected blood were consumed, stomach acid would destroy the virus.
posted by desjardins at 7:12 PM on August 18, 2008 [1 favorite]

The fact that HIV does not live long outside the body means that your hypothetical needs to contain information about how long the blood was outside the infected person's body.
posted by desjardins at 7:13 PM on August 18, 2008

Basically zero unless you have a break in your mucosa - cut in mouth, sore, esophageal tear, etc.
posted by sero_venientibus_ossa at 5:20 AM on August 19, 2008

Again, I was trying to establish a loose upper bound.
posted by jedicus at 9:58 PM on August 18 [+] [!]

Then 100% is just as good an answer as you gave, since it also ignores many important factors...

The exact answer is, as xchmp pointed out, essentially unknowable, so a guess at upper and lower bounds is as good as it gets.

The exact answer is unknowable, but that's true of all statistics. We only "know" estimates, not exact answers.

Given that bleading holes in the mouth and throat during drinking aren't as likely as friction burns in the mucous linings of the genitals, 0.67% seems like a reasonable conservative estimate (assuming that the drinker didn't just brush, floss, or bite their cheeks).
posted by IAmBroom at 12:15 PM on August 19, 2008