with a generally good chance of survival, cancer incidence (ascertainable from cancer-registration programs) is a better indicator of cancer risk than cancer-related mortality.
Results of cancer studies conducted in animal models are inconsistent (see Chapter 3). Several studies reported positive findings with respect to the development of a variety of cancers (including lung and renal cancers, leukemia, and sarcoma) in animals exposed by inhalation of uranium-ore dust or uranium dioxide, intratracheal injection of 235U (as tetravalent or hexavalent uranium), or implantation of depleted-uranium pellets (Leach et al., 1973; Filippova et al., 1978; Mitchel et al., 1999; Hahn et al., 2002; Miller et al., 2005). However, other studies reported no increase in tumor development in animals exposed by inhalation of uranium-ore dust or ingestion of uranium (Maynard and Hodge, 1949; Cross et al., 1981; ATSDR, 1999).
Twenty-three studies of uranium-processing workers examined the association between exposure to uranium and lung cancer, as did three studies of military populations and three studies of residents (see Table 8-1). Four of the uranium-processing studies reported statistically significantly increased standardized mortality ratios (SMR) (that is, above 100). All four of those studies involved the same cohort of Oak Ridge, Tennessee, and all included employees of the Y-12 plant (see Table 8-2). The specific study populations overlapped, but each study took a different approach and examined a different timeframe. The most recent study of the cohort, by Richardson and Wing (2006), did not demonstrate a statistically significant increase in lung-cancer mortality in any dose stratum. However, when assessing the dose-response relationship with a 5-year lag assumption, they found a dose-response trend between external exposure and lung-cancer mortality (due largely to a small number of excess deaths among those who accumulated an external dose of 50 mSv or more) but did not find a similar trend for internal exposure. Analyses of the joint effects of external and internal exposures found that compared to the referent group (defined as less than 10 mSv external and internal dose), the rate ratio estimates were increased for each group defined by higher cumulative concentrations of internal and/or external dose; however, the results were not statistically significant and a dose-response trend was not observed. One major limitation of the uranium-processing worker studies is the lack of control for smoking, a major risk factor for lung cancer.
Contrary to the Y-12 cohort finding, a UK study of processors found significant reductions in both mortality from lung cancer (SMR, 85; p < 0.05) and incidence of lung cancer (standardized incidence ratio [SIR], 75; p < 0.001) but is limited by having only external-exposure data (McGeoghegan and Binks, 2000b). Beral et al. (1988) also reported a significant deficit in lung-cancer mortality (SMR, 64; p < 0.01) in employees of UK atomic-weapons research establish-