 In the early morning hours of August 6, 1945, a B-29 bomber named Enola Gay took off from the island of Tennian and headed north by northwest toward Japan. The bomber's primary target was the city of Hiroshima. Hiroshima had a civilian population of almost 300,000 and was an important military center containing about 40,000 soldiers. We all know what happens next. Those closest to the bomb died instantly. Within minutes, 9 out of 10 people, half a mile or less from ground zero, were dead. The firestorm eventually engulfed 4 square miles of the city, killing anyone who had not escaped in the first minutes after the attack. Some 70,000 people probably died as a result of the initial blast, heat, and radiation effects. By the end of 1945, because of the lingering effects of radioactive fallout and other after effects, the Hiroshima death toll was probably over 100,000. The five-year death toll may have reached or even exceeded 200,000 as cancer and other long-term effects took hold. What medical lessons can we learn from atomic catastrophes such as Hiroshima, Chernobyl, and K-19? We know that the closer you are to a radiation source and the longer you remain, the higher the dose you get. The higher the dose you get, the greater your risk of health effects. We could draw a line on a graph to represent this relationship. My question to you is, can we assume that the relationship is linear across the whole range of exposures? One unit of absorb radiation is called a REM. 1000 REM is probably 100% lethal within a few weeks. 500 REMs is about 50% lethal within 30 days. Can we then assume that one REM is 0.1% lethal within a longer period? Is this, in scientific terms, a model that is linear, or at least monotonic, right down to zero radiation exposure? Is it right to determine the effects of low doses by observing the effects of high doses and extrapolating down? That is, can we make decisions about normal radiation exposure from the victims of Hiroshima? The Department of Energy initially rejected such conclusions. There were two reasons. One, the results of such a linear model with no threshold of exposure meant that every source of radiation had to be treated as a threat to human health, which was going to be economically inconvenient. And two, there wasn't much data on the health effects of low doses of radiation. There still isn't much data on this topic. It could be that the model is linear down to a single neutron or x-ray, although that seems unlikely. It could be linear down to a threshold that our bodies are capable of coping with, a safe level of radiation exposure. And it could actually be a non-monotonic U or J-shaped curve. This last is based on an idea called hormesis, that small amounts of radiation are actually good for us, because they stimulate repair mechanisms without causing much harm themselves. Here's what we do know about radiation exposure. We know that a single ionizing event can lead to DNA changes that can give rise to cancer. We know that we are constantly exposed to radiation sources, about 360 millirims per year on average in the U.S. We know that people who have a higher exposure don't necessarily have a higher risk of cancer. But we have to be careful with those kinds of associations. As I pointed out in my previous video, they don't always equal causation. We've also got some good data in live mice and human cells grown in culture, that we have a capacity to absorb and repair low level doses of radiation. So there's no obvious answer to which model best represents the health effects of radiation exposure. LNT, linear non-threshold, linear threshold or hormesis, non-monotonic curve. The linear no-threshold model is the current model used by the DOE, primarily because it is the most conservative of the three choices. But the downside is that we as a society treat the smallest exposure to radiation as an incremental risk. But this video isn't just about radiation exposure. It's about all the other ways we apply the linear non-threshold model in our lives. We make an assumption about toxic substances that isn't justified by observation. This has led to a lot of pseudoscience and irrational fear about health effects from toxins in our environment, mercury in vaccines, fluoridation in our drinking water and pesticides in our food. In some cases, alarm is warranted. There are genuine reasons to fear toxins. But we need to guard ourselves against the false assumption of extrapolating high dose effects to low dose effects. Toxicologists have a saying, the dose makes the poison. By this, they mean any substance in the proper amount is poisonous to our bodies. The measure used most frequently to express toxicity is the median lethal dose, or LD50, the dose at which 50% of a population is killed. The LD50 for sodium cyanide is 6.4 milligrams per kilogram of body weight. The LD50 for sodium chloride is 3,000 milligrams per kilogram of body weight. Even sucrose or table sugar has an LD50, 30,000 milligrams per kilogram of body weight. If we imagine a hypothetical adult weighing 80 kilograms, about 175 pounds, to be reasonably certain of killing him, I would need to give him about 480 milligrams of cyanide, about the size of a large tablet pill, or 240 grams of table salt, over half a pound. I would need 10 times that, about 5 pounds of pure cane sugar to kill him. So where does thimeros all fit into this picture, the substance that has so many people upset about health risks? Well, the LD50 is around 50 milligrams per kilogram body weight. So our hypothetical person would be killed by ingesting about 4,000 milligrams in one sitting, which would be a solid pallet about the weight of a US quarter dollar coin. Thimerosal is broken down to ethyl mercury, which rapidly clears the body with a half-life in blood, of less than four days. So unless you're taking a medium dose every day, it's unlikely to build up. This could be a problem if you eat ocean fish or shellfish every day of your life. So the LD50 is 50 milligrams per kilogram. How much is in a flu shot? 25 micrograms, or 0.025 milligrams. That's the equivalent in toxicity of 15 milligrams of table salt, about 45 grains. So if 4,000 milligrams of thimerosal kills half the population, will 1, 160,000th of that kill some subset of the population? Or is there a safe threshold of thimerosal? Anti-vaccine activists aren't really saying mercury will kill us. The concern is that it has the capability to produce autism or neurological problems, and it does. But it does so at doses far in excess of what is found in vaccines. An industrial dumping incident in Japan at a place called Minimata Bay claimed over 2,000 victims of mercury poisoning. Daily dietary intake there was about 225 milligrams of methylmercury. Methylmercury has a much longer half-life in the blood than methylmercury from thimerosal, and is therefore more toxic over time. And they were getting the equivalent of 900 flu vaccines worth of mercury every day.