Why is it dangerous to eat meat which has been left out and then cooked?
If cooking meat kills bacteria, and bacteria are responsible for problems with eating meat which has been left out, then why is it dangerous to eat meat which has been left out at room temperature and then thoroughly cooked?
A related question mentions that "Even if the bacteria is dead, toxins can remain if the food was out too long, causing problems". However, there's no further detail given. Are these toxins as dangerous as the bacteria themselves? How long does meat have to be left out to accumulate a dangerous level of toxins and thus be dangerous even if thoroughly cooked? Are these toxins the reason for the usual guideline of keeping meat unrefrigerated for a maximum of 2 hours?
The revelation, courtesy of Aaronut, that e. coli is actually dangerous because of its toxins — which cannot be denatured at temperatures which will leave meat in an edible state — has pretty much answered this question. And also given me further incentive to stop eating meat altogether :)
Our discussion (see the comments on hobodave's answer) has progressed into the realm of microbiology. Some highlights from my ongoing research:
Detail on heat-shock proteins. These seem to be the reason for the importance of keeping meat at a high temperature for a period of time.
Some background on heat resistance in bacteria. This also provides fascinating insight into how bacteria evolved immunity to antibiotics. Fungi.
I suggest also taking a look at our questions tagged food-safety. We've amassed a great number of quality Q&A on this topic.
I was asking a very similar question on Biology SE and received an awesome answer: https://biology.stackexchange.com/questions/68473/what-really-causes-the-toxicity-of-wrongly-thawed-meat
hobodave's answer is most of the way there but I think it understates the importance of protein toxins. With the vast majority of foodborne illnesses, the bacteria aren't particularly harmful at all; what you need to worry about is the protein toxins they produce.
E.Coli - probably the most well-known form of food poisoning along with Salmonella - is actually a harmless bacteria that already lives in your lower intestine. But there is a particular strain of E.Coli, notably O157:H7, that is primarily associated with food poisoning. The reason? It produces what's called a Shiga-like Toxin.
E.Coli contamination is actually dangerous on two fronts. Because the bacteria are so well-adapted to surviving in the human digestive system (as I pointed out earlier, that's their primary habitat), ingesting even a relatively small number of the bacteria will result in them multiplying and producing those toxins in your gut (and the rest of the way down). This is why it normally takes several days for you to feel the effects of this type of food poisoning; that's how long it takes for them to produce the toxins in sufficient quantity for your body to notice.
But they don't need to be in your gut to produce those toxins; a piece of meat at room temperature provides good enough conditions and more than enough raw material for them reproduce and emit those same toxins. So if you leave it sitting out too long, then it really doesn't matter how many bacteria you kill, you are going to end up with E.Coli poisoning fast, because you don't even need to wait for them to produce the toxins; they're already there.
The problem is that you can't "kill" a protein toxin with a brief burst of heat because a protein isn't alive. It's just a protein. The temperatures and times needed to destroy that toxin would be similar to the temperatures and times needed to destroy all of the protein in the food, draining all the nutrition value and quite possibly turning it into a lump of charcoal.
Salmonella seems to be a fountain of misinformation with all sorts of people saying that it doesn't produce toxins. This simply isn't true. Inside the host it produces what's called an AvrA toxin (which isn't "toxic" per se, but allows the bacteria to grow to larger numbers), and some strains can also produce a CdtB toxin, which is highly toxic. (Apparently there's also a similar toxin produced by other strains.) I'll be honest, a lot of the medical mumbo-jumbo is way beyond my ability to comprehend, but it seems that a lot of the public confusion comes from the fact that salmonella can do some nasty things even without the toxins - but that doesn't mean that the toxins themselves can't do plenty of damage even if you manage to kill the bacteria.
The same applies to many other types of dangerous bacteria; C.diptheriae produce the diphtheria toxin, C.botulinum produce the botulinum toxin (botulism); even the infamous mad cow disease was, as far as we know, caused by a protein, not a bacteria, which is why it was able to be transmitted to humans even through cooked beef.
Are protein toxins the only reason why the USDA insists on a maximum 4-hour cumulative danger zone? Probably not. As hobodave says, the more the bacteria multiply, the harder is to kill all of them, even at high temperatures. The figure of 74° C / 165° F that the food agencies give us for poultry is not going to kill exactly 100% of all the bacteria, and if it only kills - I'm just throwing out a number here - 99.999% of them, that may be good enough for relatively fresh poultry but won't be enough if you've got a whole bacterial colony to worry about.
We can only speculate as to exactly what's entailed by the "danger zone" but my guess is that it's actually a combination of statistics, probabilities, and safety margins, which include, but are not limited to, the effects of protein toxins.
Very informative, thank you. So I guess that while the cow is still alive, its antibodies would be limiting the numbers of *e. coli* and cleaning out the toxins that it is immune to? Otherwise, there wouldn't be any difference between leaving the meat out at room temperature and leaving the cow out at room temperature before slaughtering it. —————— Also, see my comments on hobodave's answer for some more questions about the 99.9*% per amount of time kill ratio at a given temperature.
@intuited: I really don't know what you're getting at. Most bacterial contamination does not come from the animal itself, it comes from the processing. If you're running your own farm and slaughtering your own animals then you don't really have to worry about a lot of this.
Also @intuited, do keep in mind that bacteria grow exponentially (at least until they run out of food), so you would need to exponentially increase your cooking times as well - assuming that were actually enough (which it likely isn't).
@Aaronut: If heat really does kill a given percentage (e.g. 99.9999%) of all bacteria every so many seconds, then this implies a direct relationship between the time bacteria have to grow and the time it takes to kill them off. E.G. suppose that cooking for 2 seconds kills 10% of the bacteria, and that bacteria multiply tenfold every minute. If there are 10 bacteria when I take the meat out of the fridge, then after 6 minutes there will be 10⁶ bacteria. If there are 10⁶ when I start cooking, then after 2 seconds there will be 10⁵; 2 seconds after that there will be 10⁴; and so on.
@Aaronut: That's assuming that the surviving bacteria are not doing so due to higher heat resistance.
@Aaronut: I don't have to worry about it unless there's a possibility that my living cattle have been exposed to *e. coli*, which they can carry without symptoms.
@intuited: The math here is not as simple as you seem to be suggesting. See Thermal death time for how this actually works. And keep in mind that there are many heat-resistant strains of bacteria that can survive far longer than 2 seconds (up to 2 minutes). There are also multiple-site and multiple-hit/multiple-target theories suggesting that is quite a bit more complicated than even the relatively simple aforementioned model.
@Aaronut: regarding the percentage of bacteria killed. You had mentioned that you were "just throwing out a number" of 99.999%. The USDA tables I've seen have indicated 6–7 decimal reduction, meaning 99.9999% – 99.99999% reduction in pathogens. Typically 6D is used for listeria and E. Coli, 7D for Salmonella
Just double-checked on this: "The 1999 FSIS final rule, Performance Standards for the Production of Certain Meat and Poultry Products, requires a 6.5 log10 relative reduction (6.5 log10 lethality) of Salmonella for cooked beef, roast beef and corned beef (9 CFR318.17)." See http://edocket.access.gpo.gov/cfr_2008/janqtr/pdf/9cfr318.18.pdf