Uranium, in particular depleted uranium, is fast acquiring notoriety as a radiological hazard. In fact, its radiotoxicity is known to be low. However, its chemical toxicity should not be ignored.

Uranium is the heaviest naturally occurring element and is found at an average concentration of 0·0004% in the earth's crust. The human body contains about 100 µg of uranium, which is derived mostly from uranium in food (daily intake about 1·5 µg), especially from vegetables, cereals, and table salt. Uranium of natural isotopic composition consists of three isotopes--U-238 (half-life [T 1/2 ]=4·47 x 109 years, 99·3% by mass), U-235 (T 1/2 =7·04 x 108 years, 0·72% by mass), U-234 (T 1/2=2·46 x 105 years, 0·006% by mass)--all of which are radioactive. Of these, only U-235 is a valuable fuel for the production of electricity. Consequently, during the manufacture of nuclear fuel the concentration of U-235 is commonly increased. The byproduct of this enrichment process is depleted uranium, which typically contains about 70% less U-235 and 80% less U-234 than does natural uranium.1 In addition, because some reactor-irradiated uranium is used for enrichment, depleted uranium may contain traces of U-236 (<0·003%) and inconsequential concentrations of other actinide isotopes, including Pu-239. Over time, a significant mass of depleted uranium has been produced and accumulated by nuclear-fuel manufacturers.

The major intended use of depleted uranium was as a cladding material in fast-breeder reactors, where its interactions with neutrons would produce additional reactor fuel, in the form of Pu-239. However, most breeder reactor programmes have now closed and other uses have been found, many of which exploit the physical properties of this metal. Depleted uranium has replaced uranium of natural isotopic composition as a less radioactive substitute for some uses--eg, laboratory chemicals. It is also used as a fluorescent additive in dental porcelain crowns, as X-ray shielding in hospitals, as counterweights for control surfaces (eg, rudders, flaps) in commercial aircraft, in the manufacture of keels for yachts, and as the armour-piercing component of some munitions. In the past uranium has also been used to colour glasses and ceramic glazes and as a pharmaceutical for the treatment of diabetes. In total, human experience of the use and toxicity of uranium or depleted uranium spans more than 20  years.

Exposure to radiations emitted by uranium metal presents a negligible radiological hazard. Completely surrounding a worker with depleted uranium for 8 h a day for a year would not result in radiation doses that exceed the maximum annual occupational dose limit for radiation workers. Similarly, if uranium is retained indefinitely in contact with the skin, the dose is not large enough to produce tissue damage.2 Consequently, only the possibility of hazards arising from depleted uranium that has entered the body needs further consideration.

In its most common, and bioavailable, forms uranium exists as a uranyl (UO22+) ion and shares many chemical and biological properties with the alkaline earth ions. 3 Therefore, after its entry into the bloodstream, either following the dissolution of inhaled uranium-containing dusts or from ingested materials, some is co-deposited on bone surfaces with calcium, where it may remain for many years. Nevertheless, about 90% is excreted in urine within 24 h after intake, with most of the remainder being excreted within the following weeks and only a few percent being retained in the skeleton. Normal excretion levels of uranium, for members of the general public, are variable but usually lie in the range 0·04-0·5 µg/L of urine. This continuous excretion of endogenous uranium complicates bioassay, so that past intakes of depleted uranium can be detected in urine only by measuring the small perturbations in the normal isotopic ratio produced by the release of depleted uranium from the skeleton or lungs. Uranium does not deposit in any other skeletal or nervous tissues and, like radium, is absent from the bone marrow .

A high level of urinary excretion of uranium is, however, associated with temporary accumulations of this metal in the kidneys. These accumulations occur after high intakes (>70-100 µg/kg body weight) and produce chemical damage to the proximal renal tubules. This damage is manifested as increased catalase excretion and proteinuria.4 At even higher levels of intake, studies in rats indicate that damage may also occur in renal glomeruli. In human beings who have had large acute intakes of uranium, these effects were transitory and wholly reversible. Moreover, the regenerated tubular epithelial cells are reported to be more resistant than the orginial cells to uranium damage, so that workers exposed frequently may have increased tolerance of high uranium intakes. 5

Some uranium is present in semen; this is not unexpected since semen, like all other body fluids, contains calcium. However, uranium is not deposited in either the testis or ovary, and genotoxic effects of uranium have not been described in either animals or man.

Although uranium binds strongly to nucleic acids in vitro and is used to stain tissue sections for electronmicroscopy, no chemical mutagenicity has been demonstrated for this element in mammals. By contrast, radiation is mutagenic. Nevertheless, experience supports the claim that health effects produced by the radioactive decay of depleted uranium in the body are extremely unlikely--a feature of its very low specific activity (specific activity of uranium isotopes being about 15 Bq/mg). Depleted uranium is about 3 million times less radioactive than Ra-226 (still found in many old luminous clocks and watches) and 10 million times less radioactive than Am-241 (found in domestic fire detectors). By contrast, high specific-activity uranium isotopes (eg, U-232 and U-233) are quite radiotoxic. It follows that uranium isotopes differ widely in toxicity, with chemical toxicity becoming relatively more important as the radiotoxicity of the isotopes decreases in the order U-232, U-233, U-234, U-236, U-235, U-238. The situation in the lungs is slightly different, since only dissolved uranium is chemically toxic. Soluble uranium is largely absent from the lungs because when formed from inhaled deposits it is rapidly transferred to the bloodstream. That insoluble uranium deposits present no immediate health hazard is evidenced by the lack of health effects in the 22 American servicemen who retain depleted uranium shrapnel in their bodies, or in the 13 suspected of retaining such shrapnel, as the result of their involvement in friendly-fire incidents during the Gulf War. Also, these servicemen have fathered more than 20 normal children.

Evidence of the low radiotoxicity of depleted uranium comes from considerations of radiation dose, epidemiological studies of uranium workers, uranium toxicity studies in animals, and human experience with radium isotopes.

The International Commission on Radiological Protection, whose recommendations are accepted globally as a valid base for radiological protection dosimetry, has published methods for calculating radiation dose to radiation-sensitive tissues and organs following intakes of radionuclides. Risk estimates, based on the known effects of radiation, can then be used to predict tumour frequency. The ICRP recommends a dosimetry model for uranium isotopes,6 which was developed over many years, from metabolic data extracted from published work on the distribution and biokinetics of uranium in man, including that provided by human injection studies. For depleted uranium the integrated, 50-year doses calculated using this model are very low (about 25 nSv/µg uptake), and uptakes of more than 5 g into the blood are needed to give a radiation dose equivalent to that received over a period of 50 years from natural, UK, background radiation. It follows that the predicted tumour frequencies are also very low for any conceivable intake of depleted uranium. For example, the risk of a fatal cancer to an operator permanently located 50 m from a gunnery target, based on 12 tonnes of depleted uranium fired over a 15-year period, was recently calculated to be 1 in a million, from an annual radiation dose of 20 mSv (for comparison, the annual background is 2200 mSv/year). 7

The specific effects of uranium can be studied in laboratory animals without interference from confounding factors such as external -radiation, radon exposures, and cigarette smoke. Moreover, such studies can reveal the latent radiotoxicity of the element by examining the effects of high-specific-activity U-233 or very large intakes of natural uranium. In rats that had inhaled natural uranium at concentrations of up to 50 mg/L for 273 h (giving radiation doses to the lung of up to 1·64 Gy), 31% developed lung tumours. In mice injected with sufficient U-233 to produce skeletal radiation doses of 2 Gy, small numbers of excess bone tumours, myeloid leukaemia, and kidney tumours were found.8 However, the number of tumours found was lower than that produced by Pu-239 and Am-241 at the same skeletal radiation dose, and in both studies the tumour frequency was consistent with the -radiation dose received by the tissues. In view of the low specific activity of depleted uranium, occupational exposures to this form of the element are extremely unlikely to result in radiation doses to tissues that are large enough to produce malignant changes.

About a dozen health-effect studies of uranium workers have been conducted. The largest are an American study of 18 869 workers, employed between 1943 and 1947 and followed up until 1974, at uranium plants in Oak Ridge, Tennessee9 and a British study of morbidity and mortality in 19 454 workers employed at the Spingfields uranium plant between 1946 and 1995.10 The American workers were exposed to high average air concentrations of uranium dust (up to 0·5 mg/L of air). In this population only the standardised mortality ratio (SMR) for lung cancer was raised (SMR=1·22, 95% CI 1·10-1·36), but this finding may have been an effect of smoking, since risk was similar in all worker groups irrespective of uranium exposure history. No excess bone cancer, leukaemia, or other cause of death was found. Similarly, in the British study no excess mortality, from any cause, was found that could be attributed to uranium exposure. Indeed, the standardised registration ratios (SRRs) for cancer were identical (0·81) for workers exposed or unexposed to uranium. The corresponding SMRs were 0·84 and 0·98, respectively. There was an excess of non-Hodgkin's lymphoma and Hodgkin's disease (SRR 0·79-2·83), but the occurrence of the disease did not correlate with exposures to uranium. Neither the American, nor the British, nor any other large study has shown an excess incidence of either bone cancer or leukaemia in uranium workers, although several showed excess lung cancer, and in one11 excess central-nervous- system cancer was indicated, but without dose response. In no study has excess cancer been linked to intakes of uranium. In short, despite numerous workers being exposed to large amounts of uranium, no study has provided evidence that either depleted or natural uranium is carcinogenic, even for lung cancer following inhalation of uranium.12

Since the uranyl ion behaves like an alkaline earth ion in the body, the toxic effects of uranium body-burdens can also be forecast by use of the extensive database on the toxicity of -emitting radium isotopes--in particular, the data for occupational exposures to Ra-226. Apart from epithelial tumours produced by the accumulation of radon gas in head sinuses, the only tumours that are produced by Ra-226 in man are bone tumours. Moreover, these occurred only in those people in whom the skeletal dose exceeded 10 Gy. No excess leukaemia or other tumours were found. Skeletal doses similar to those accumulated by the radium workers cannot be reached after industrial or occupational intakes of either depleted or natural uranium. Even if they could, the only tumours expected would be bone tumours and, in the case of inhaled deposits, lung tumours.

It can be safely concluded that at any conceivable level of uptake depleted uranium will have no appreciable radiological or chemical carcinogenic potential. Moreover, even if cancers were to be produced, they would occur many years after intake, because of the lag period between damage to sensitive cells and the appearance of recognisable tumours. In man, for chronic irradiation from an internally deposited -emitting radionuclide, these latency periods would typically lie in the range of 10 years to several decades. In view of this latency, tumours in individuals exposed for shorter periods--eg, in servicemen exposed to depleted uranium in the former Yugoslavia within the past decade--cannot be attributed to radiation from depleted uranium. Finally, the only chemical toxic effect expected would be reversible damage to the kidney.

N D Priest

School of Health, Biological and Environmental Sciences, Middlesex University, London, N11 2NQ. UK (e-mail:n.priest@mdx.ac.uk)

2 Fetter S. and von Hippl FN. The hazard posed by depleted uranium munitions.Sci Glob Security 1999; 8: 125-61. [PubMed]

3 Priest ND. The distribution and behaviour of metals in the skeleton and body: studies with bone seeking radionuclides. In: Priest ND, Van de Vyver F, eds. Trace metals and fluoride in bones and teeth. Boca Raton: CRC Press, 1990: 83-140.

4 Friberg L, Nordberg GF, Vouk VB. Handbook on the toxicology of metals, volume II. Elsevier: Amsterdam, 1986: 623-37.

5 Hodge HC, Stannard JN, Hursh JB. Handbook of experimental pharmacology, vol. 36, uranium, plutonium, transplutonic elements. New York: Springer-Verlag, 1973: 165-95.

6 Internationl Commission on Radiation Protection. Age-dependent doses to members of the public from intake of radionuclides: part 3. ICRP publication 69.Ann ICRP 1994; 25: 57-74. [PubMed]

7 Westlakes Research Institute, Cumbria. Press release, Jan 12, 2001.

8 Ellender M, Harrison JD, Pottinger H, Thomas JM. Induction of osteosarcoma and acute myeloid leukaemia in CBA/H mice by the alpha-emitting nuclides, uranium-233, plutonium-239 and americium-241.Int J Radiat Biol 2001; 77: 41-52. [PubMed]

9 Polednak AP, Frome EL. Mortality among men employed between 1943 and 1947 at a uranium-processing plant.J Occup Med 1981; 23: 169-78. [PubMed]

10 McGeoghegan D, Spinks K. The mortality and cancer morbidity experience of workers at Springfields uranium production facility, 1946-95.J Radiol Protect 2000; 20: 111-37. [PubMed]

11 Checkoway H. Pearce N, Crawford-Brown DJ, Cragle DL. Radiation doses and cause-specific mortality among workers at a nuclear materials fabrication plant.Am J Epidemiol 1988; 127: 255-66. [PubMed]

12 Ritz B. Radiation exposure and cancer mortality in uranium processing workers. Epidemiology 1999; 10: 531-38. [PubMed]