The Myth of Plutonium Toxicity

Bernard L. Cohen

Plutonium is constantly referred to by the news media as "the most toxic substance known to man." Ralph Nader has said that a pound of plutonium could cause 8 billion cancers, and former Senator Ribicoff has said that a single particle of plutonium inhaled into the lung can cause cancer. There is no scientific basis for any of these statements, as I have shown in a paper in the refereed scientific journal Health Physics (Vol. 32, pp. 359-379, 1977). Nader asked the Nuclear Regulatory Commission to evaluate my paper, which they did in considerable depth and detail, but when they gave it a "clean bill of health" he ignored their report. When he accuses me of "trying to detoxify plutonium with a pen," I offered to eat as much plutonium as he would eat of caffeine, which my paper shows is comparably dangerous, or given reasonable TV coverage, to personally inhale 1000 times as much plutonium as he says would be fatal, or in response to former Senator Ribicoff's statement to inhale 1000 particles of plutonium of any size that can be suspended in air. My offer was made to all major TV networks but there has never been a reply beyond a request for a copy of my paper. Yet the false statements continue in the news media and surely 95% of the public accept them as fact although virtually no one in the radiation health scientific community gives them credence. We have here a complete breakdown in communication between the scientific community and the news media, and an unprecedented display of irresponsibility by the latter. One must also question the ethics of Nader and Ribicoff; I have sent them my papers and written them personal letters, but I have never received a reply.

Let's get at the truth here about plutonium toxicity. We begin by outlining a calculation of the cancer risk from intake of plutonium (we refer to it by its chemical symbol, Pu) based on standard procedures recommended by all national and international organizations charged with responsibility in this area, and accepted by the vast majority of radio-biomedical scientists.

Estimate of Plutonium Toxicity from Standard Procedures

The first step is to calculate the radiation dose in rem (the unit of dose) to each organ of the human body per gram of Pu intake. According to ICRP (International Commission on Radiation Protection) Publication No. 19, about 25% of inhaled particles of the size of interest (0.5-5 micrometer in diameter) deposit in the lung, and 60% of this is eliminated only with a 500-day half-life. From this information and the known rate and energy of [alpha]-particle emission, we can calculate the radiation energy deposited in the lung, which is directly convertible to dose in rem.

According to ICRP Publication 19, 5% of inhaled Pu gets into the bloodstream from which 45% gets into the bone and an equal amount collects in the liver; the times required for elimination from these are 70 and 35 years, respectively. This is all the information needed to calculate doses to bone and liver in rem per gram of Pu inhaled.

If Pu is ingested with food or water in soluble form, the ICRP estimates that 3 x 10^-5 (30 parts per million) gets through the intestine walls into the bloodstream. From this and the information given above, calculation of rem to the bone and liver per gram of Pu ingested is straightforward. In addition, there is dosage to the gastrointestinal tract calculable by ICRP prescriptions.

Once the dose in rem is calculated, the next step is to convert this to cancer risk using the BEIR Report, the standard reference in this area produced by the National Academy of Sciences Committee on Biological Effects of Ionizing Radiation. It recommends a model in which there is a 15-year latent period following exposure during which there are no effects, followed by a 30-year "plateau" period during which there is a constant risk of 1.3 x 10^-6 (1.3 chances per million) per year per rem for lung cancer and 0.2, 1.0, and 0.3 x 10^-6 per year per rem for bone, gastrointestinal tract and liver* cancer, respectively. For children less than 10 years old, these are divided by five, and for an older person, there is a calculable probability that death will result from other causes before the cancer develops. With this information we can calculate the cancer risk as a function of age at intake. Averaging over ages, we obtain the average cancer risk per gram of Pu intake. * In the BEIR Report, liver cancer is included among "all other" for which the risk is 1.0 x 10^-6, the value used here is based partly on other information.


Table I: Cancer Doses in Micrograms

Entrance Mode



Inhalation (dust in air)



Ingestion with food or water

6.5 x 10^6


The results are given in Table I for the most important isotope of Pu, 239-Pu, which contains 1 curie of radioactivity for each 16g, and for the mixture of Pu isotopes that would be commonly found in power reactors, which is 6 times more intensely radioactive (1 curie in each 2.5 g). We refer to the latter as "reactor-Pu" and use it in our discussions where appropriate.

Table I shows the inverse of the risk, which we call the "cancer dose." For example, we see that the risk of inhaling reactor-Pu is 1/200 per [micro]g, so if one inhales 10 [micro]g, he has one chance in 20 of developing cancer as a result. Another application is that in a large population we may expect one cancer for every 200 [micro]g inhaled, so if a total of 1000 [micro]g is inhaled by people, we may expect 5 cancers (regardless of the number of people involved).

Estimates of cancer doses of Pu have also been derived using different methods by the British Medical Research Council in its report "The Toxicity of Plutonium," and by Dr. C.W. Mays (who developed some of the important basic information in his experiments on dogs) in a report published by the International Atomic Energy Agency (IAEA-SM-202/806), and they agree closely with Table I. We see from Table I that Pu is dangerous principally when inhaled as a fine dust. It is not very toxic when ingested with food or drink because of its very small probability of passing through the intestine walls into the bloodstream. Pu forms large molecules, which have great difficulty in passing through membranes.

In addition to causing cancer, intake of plutonium can also cause genetic defects among progeny in the next 5-10 generations, but the total number of eventual genetic defects before they are bred out is only 20% of the number of cancers. For simplicity we restrict our discussion to cancers, but the genetic effects can always be included by applying the 20% addition.

The estimates in Table I are based on data from radiation effects on humans as analyzed in the BEIR Report. These include Japanese A-bomb survivors, miners exposed to radon gas, people treated for various maladies with radium or with X-rays, etc. None of these effects were from Pu -- there is no evidence for any injury to humans from Pu toxicity. However, there is a considerable amount of data from animal studies with Pu, and this is summarized for lung cancer in Fig. 1 where the line shows the estimate from our calculation. In general the agreement is quite reasonable.

[Omitted: "FIGURE 1. Data from animal studies with Pu, summarized for lung cancer." The graph shows 40 data-points, with confidence intervals, from animal studies (dogs, mice, rats, rabbits) with a calculated line over them. The x-axis, which is logarithmically scaled, is labeled "Dose to Lung (millions of millirem)" and the y-axis is labeled "Incidence of Lung Cancer (%)". Taking representative points from the calculated line in the figure, we get: (~0.3 Mmrem, ~1%), (~1.0 Mmrem, ~5%), (~10.0 Mmrem, ~38%), (~11.0 Mmrem, ~65%). Mmrem: millions of millirem.]

There has been a great deal of publicity about the high point for beagle dogs (the highest point in Fig. 1) but we see that our curve passes within the error bars given by the authors. One aspect of the experiment that is frequently overlooked is that the latent period for development of the cancers increased with decreasing dose, and in fact the dogs contributing to the point under discussion developed cancer rather late in life. If this effect is extrapolated to lower doses, the latent period for most doses usually considered would greatly exceed life expectancy, so the effects we derive in this paper would be substantially reduced.

Criticisms of Standard Procedures

There have been several criticisms of treatments like the one we have given. The best known of these is the "hot-particle" theory, which gives greatly increased effects (by a factor of 100,000) due to the fact that the Pu is not evenly distributed over the lung but is concentrated in particles, which give much higher than average doses to a few cells. This theory has been studied and rejected by the following groups:

  • A Committee of the U.S. National Academy of Sciences especially assembled for this study in a report entitled "Health Effects of Alpha-emitting Particles in the Respiratory Tract

  • U.S. National Council on Radiation Protection and Measurement (NCRP), a very distinguished group composed of about 70 in our nation's leading radio-biomedical research scientists, in NCRP Publication No. 46

  • British Medical Research Council in "The Toxicity of Plutonium"

  • U.K. National Radiological Protection Board in its Report R-29 and Bulletin No. 8 (1974)

  • U.S. AEC in a very elaborate study, WASH-1320, authored by three of the world's leading researchers on Pu toxicity

  • U.S. NRC in Federal Register, Vol. 41, No. 76

  • U.K. Royal Commission on Environmental Pollution -- Sixth Report -- Nuclear Power and the Environment

One easily understood aspect of these criticisms is that there were about 25 workers at Los Alamos who inhaled varying amounts of Pu about 30 years ago, and according to the "hot-particle" theory each should have experienced about 200 lung cancers, whereas there have been no lung cancers as yet among them. According to our estimates in Table I, there is a 40% chance that one of them would have had lung cancer, so this is experimental evidence that Table I does not grossly underestimate the cancer risk from Pu intake. [For more on the Los Alamos workers see George L. Voelz, Robert S. Grier, Louis H. Hempelmann, "A 37-Year Medical Follow-Up Of Manhattan Project Pu Workers", Health Physics, Vol. 48, No. 3 (March 1985), pp. 249-259.]

Another criticism of the "hot-particle" theory is that there are experiments on animals in which two groups were exposed to the same total amount of Pu but in one of them it was much more in the form of hot particles -- and that group experienced fewer cancers. It was also pointed out that particles in the lung do not stay in one place but are constantly moving about so that their exposure does not fall on only a few cells.

After these rejections of the "hot-particle" theory appeared, John Gofman, a former research scientist who has spent the past several years as the full-time leader of an anti-nuclear organization, came out with a new theory ascribing enhanced toxicity to Pu. His paper was not written for a scientific journal but was inserted in the congressional Record by Senator Gravel. His basic premise was that smoking destroys the cilia, the fine hairs that stop dust particles from entering the bronchial region -- this much was well established -- and that Pu particles therefore remain in that region for a very long time, allowing their radiation to cause bronchial cancers. This allows him to ignore the animal data as animals do not smoke. He also manages to explain the lack of lung cancers among the 25 Los Alamos workers by a combination of four improbable hypotheses, the failure of any one of which would destroy his theory.

There have been at least seven individual critiques of the Gofman theory. Perhaps the most telling criticism is that there was a series of experiments at New York University in which a number of graduate students inhaled a controlled amount of radioactive dust and the rate at which this dust was cleared from the bronchial region was directly determined by placing radiation detectors over their chests and measuring the radiation intensity as a function of time. It was found that there was no difference between smokers and nonsmokers, and the experimenters concluded that smokers do more coughing and have increases mucus flow, which compensates for their lack of cilia. In fact, if dust accumulated in the bronchial region of smokers in the manner postulated by Gofman, their bronchial tubes would be completely closed and they would die by suffocation.

There were many more weak points in the details of the Gofman paper. He misuses the BEIR Report, he miscalculates the area of the bronchial region by a factor of 17 and thereby incorrectly increases the toxicity by that factor, he misuses the ICRP lung model, etc. He even suggests that the great increase in lung cancer in recent years may be due to Pu, but this increase has been steady since the 1930s whereas Pu-induced cancers should not have occurred until 1960. Moreover, the lung cancer increases have been in areas with chemical industry and high air pollution, and there has been no increase in areas downwind from the Nevada test site where Pu would have its maximum effect.

A relatively less publicized attack on the conventional approach to evaluating Pu toxicity is the "warm-particle" theory of Edward Martell. He hypothesizes that natural radiation is one of the principal causes of lung cancer, but this idea has not been accepted by the cancer research community.

K.Z. Morgan has proposed that the relative biological effectiveness (RBE) for Pu in bone might be 250 times larger than the usual value. C.W. Mays, on whose experiments much of Morgan's hypothesis is based, reanalyzed Morgan's work and concluded that if his approach is correct, the increase should be only by a factor of 10. There is experimental information on this from some supposedly "terminally ill" patients injected with Pu in 1945-46 to study Pu metabolism. Four of these are still alive and one who was injected with a rather large quantity died of unrelated causes only in 1968. If the RBE of Pu were 10 times the present value, there is a better than even chance that one of these five would have gotten bone cancer, but none did. As our calculated inhalation effects are dominated by lung cancer, a factor of 10 increase in bone cancer risk would only double the total inhalation risk.

S.M. Wolfe, and employee of a Nader-sponsored group, drew far-reaching conclusions from the fact that 11 of the first 30 deaths in the US Transuranic Registry (a registry of people who have worked with plutonium) revealed cancers on autopsy, whereas based on listed cause-of-death for all U.S. males, only 6.2 of each 30 deaths is from cancer. His paper, which was never published in the scientific literature, received very wide publicity in the news media. However, it turned out that autopsies were done preferentially on people who had died of cancer, and that explained the entire effect. In addition, it was pointed out that Pu is expected to cause cancers of the lung, bone, and liver, whereas among the 11 cases there were no bone or liver cancers, and less than the expected number of lung cancers for a normal population. Needless to say, the news media never bothered to report that the Wolf paper was based on an incorrect premise.

In evaluating all of the criticisms outlined above, it is important to realize that they are actively considered every year by a committee of the ICRP and that they have repeatedly been rejected. Likewise, the EPA, which has jurisdiction in the U.S., studied the matter and decided not to modify its standards. No standard-setting or official study group in any country has given credence to any of these criticisms of the standard procedures we used in deriving Table I.

Consequences of Plutonium Dispersal

It is clear from Table I that Pu is dangerous principally as an inhalant, so we now consider the consequences of a dispersal of Pu powder in a populated area. The calculations are done with a standard meteorological model, in which the dust cloud moves with the wind dispersing in the downwind, crosswind, and vertical directions. Meteorologists have determined the extent of dispersal as a function of wind velocity and atmospheric stability. Figure 2 shows the results of calculations assigning the atmospheric stability most characteristic of each wind velocity. This is different between day and night, so separate curves are given for each.

These curves give the area within which various fractions, q/Q, of the dispersed Pu are taken in by a person inhaling at an average rate. For example, we see from Fig. 2 that for a typical daytime 8 m/sec wind velocity, only in an area of 500 m^2 is as much as 10^-6 (one millionth) of the dispersed Pu inhaled. A typical city population is 10^-2 people/m^2, so there would typically be about 5 people in this area. Similarly, from Fig. 2, about 60 people would inhale 10^-7, 700 people would inhale 10^-8, etc. of the dispersed Pu.

As we know the cancer risk per microgram of Pu inhaled from Table I, it is straightforward to calculate the total number of cancers expected per gram of Pu dispersed. When corrections are applied for the fraction of typical Pu powders that are in particles of respirable size, the efficiency of dispersal, the protection afforded by being inside buildings, and decreased breathing rates at night, the result is that we may expect about one eventual cancer for every 24 g of Pu dispersed, or about 19 fatalities per pound.

If there is a warning, as in a blackmail scenario, people can be instructed to breathe through a folded handkerchief or a thick article of clothing, with a resulting decrease in fatalities to 3 per pound dispersed.

Eventually, the Pu settles to the ground but it may then be blown up by winds. Meteorologists have also developed methods for calculating these effects ("deposition" and "resuspension"). Within the first few months, this causes about one-third as many cancers as inhalation from the initial cloud. Beyond this time period, resuspension is of much less and continually decreasing importance as the Pu becomes part of the soil.

[Omitted: "FIGURE 2. Area over which the ratio of inhaled to dispersed Pu has values shown for q/Q versus wind velocity under typical day and night atmospheric conditions." The x-axis, which is logarithmically scaled, is labeled "Wind Velocity (meters/sec)" and the y-axis is labeled "Area (meter^2)".]

Of course, 239-Pu lasts for tens of thousands of years, so let us consider its effects over this time period. We know the amount of uranium in soil and we know now how much there is as dust in the air, so we can estimate how much is inhaled per year -- it calculates out to be 1.3 x 10^-11 of that in the top 20 cm of soil. If this factor is applied to the Pu after it becomes part of the soil, we find that over the 25,000-year half-life there will eventually be about one fatality per 2500 g of Pu dispersed. Thus, we see that the long half-life is almost irrelevant; nearly all of the damage eventually done occurs very soon after dispersal.

A summary of all these effects of Pu dispersal is given in Table II. It also includes plant uptake into food. There is a great deal of information on uptake of Pu by plants both from laboratory experiments and from several areas where an appreciable amount of Pu has gotten into the soil from bomb tests or from various research activities. Plant uptake is small for the same reason that Pu does not easily pass through the walls of the intestines -- it forms large molecules, which do not easily pass through membranes. From Table II we see that the total eventual effect of Pu dispersal in a city is one fatality per 18 g dispersed without warning, or 25 fatalities per pound.


Table II: Fatalities per Gram of Reactor-Pu Dispersed

Inhalation from cloud

0.042 (1/24)



Long Term

0.0004 (1/2500)

Plant uptake into food



0.058 (1/18)

Dangers of Plutonium Dispersal

The fear is sometimes expressed that the world may become "contaminated" with 239-Pu. To evaluate this potentiality, we calculate that if all the world's present electric power were produced by fast breeder reactors in an equilibrium situation where Pu is consumed as fast as it is produced, the total amount of 239-Pu existing in the world would be 2 x 10^8 curies.

By comparison, the radium (226-Ra) in each meter of depth of the earth's crust is 1.2 x 10^9 curies, so there is as much Ra in each 17 cm of depth as there would be 239-Pu in the whole world. For ingestion, Ra is 40 times more toxic than Pu as it passes through the intestine walls much more easily. For direct inhalation, Ra is less hazardous than Pu, but it serves as a source of radon gas, which comes up out of the ground and mixes with the air we breathe, and therefore is a serious inhalation hazard, so as material on the ground, Ra is a 40-fold greater inhalation hazard than Pu.

Thus, as a long-term hazard either for ingestion or for inhalation, Ra is 40 times worse than Pu; the total Pu in existence for an all-breeder power system would then be as dangerous as the Ra in each 4 mm of our soil. Of course, nearly all of this Pu would be in reactors or in other parts of the nuclear industry, well isolated from the environment.

There is now a legal requirement on the allowable releases of Pu from nuclear plants, which is such that if all U.S. power were nuclear and derived from fast breeder reactors (they use the most Pu), the total releases would be about 0.6 g/year. If we use table II, this would predict an average of 0.03 fatality/year, but that would be valid only if nuclear plants were in cities; as they are not, the expected effects are about 10 times less, or one fatality in 300 years.

Some perspective on this problem may be obtained by comparing the 0.6 g/year that may some day be released by the nuclear industry with the amount of Pu that has been dispersed in the atmosphere in nuclear bomb tests, which is 5 million g. Estimates on the same basis that we have been using predict about 200 U.S. fatalities to date from Pu releases in bomb tests, and 4000 in the world. It also predicts about 200 fatalities worldwide from the reentry burn-up in 1964 of a space vehicle carrying a SNAP-9A 238-Pu-powered energy source. It is important to keep in mind that all of these estimates are theoretical. These is no direct evidence for Pu toxicity having caused serious injury to any human being, anywhere, ever.

The reason why the legal requirement on plutonium releases is so stringent is not because Pu is so dangerous, but because the technology is available for keeping the releases that low, and in fact this technology is very close to present practice. Pu dust particles tend to stick to each other and their containers, so Pu is not easily dispersed. It is also very readily collected on filters; anywhere Pu powder is used, the air is exhausted through filters, which catch all but about one part per billion of the dust suspended in air.

Of course, the control measures are expensive and they increase the cost of nuclear electricity. As previously noted, the reason they are required is not because Pu is so dangerous -- one fatality every 300 years is surely a trivial problem when burning coal, our only viable alternative to nuclear energy, is killing 10,000 people every year with its air pollution -- but because the public is afraid of plutonium. Ralph Nader, former Senator Ribicoff, John Gofman, and their like have done their work well, and the public is paying the price in its electric bills.

One often hears that in large-scale production of Pu we will be creating unprecedented quantities of a poisonous material. Because Pu is dangerous principally as an inhalant, we compare it in Table III with quantities of other poisonous inhalants produced in the U.S. We see that it is relatively trivial by comparison. Moreover, it should be noted that Pu is not easily dispersed whereas the others are gases and hence readily dispersible. Of course, Pu released to the environment will last far longer than these gases, which would be decomposed chemically, but recall from our earlier discussion that nearly all of the damage done in Pu dispersal is by the initial cloud of dust; all of the later resuspension and the thousands of years spent in the soil do far less damage. It is thus not unfair to compare Pu with the poison gases, and we see from Table III that it will always be far less of a hazard.


Table III: Lethal Inhalation Doses Produced Annually in the U.S. (x 10^12)







Hydrogen cyanide


Pu if all U.S. power were from fast breeder reactors


It is often argued that there is a great deal we do not know about Pu toxicity. While this may be true, one would be hard-pressed to name another public health issue that is as well understood and controlled. Surely it would not be air pollution from burning coal, which is a million times more serious a problem. Surely it is not food additives or insecticides or such [the dangers from these have also been greatly exaggerated] that may well be doing real harm to our health. Pu hazards are far better understood than any of these, and the one fatality per 300 years they may someday cause is truly trivial by comparison.

In spite of the facts we have cited here, facts well known in the scientific community, the myth of Pu toxicity lingers on. The news media ignore us, and prefer to continue scaring the public at every opportunity. They don't recognize the difference between political issues on which everyone is equally entitled to an opinion, and scientific issues, which are susceptible to scientific investigation and proof. The myth may linger forever.



From Nuclear Energy, Karl Otto Ott and Bernard I. Spinard, eds., New York: Plenum Press, 1985, pp. 355-365.

Add new comment