A large multi-country case-control study of arsenic and cancer risk was completed, based on multi-disciplinary work in the fields of epidemiology, toxicology, analytical chemistry and genetics. A total of 948 cancers and 540 controls were included in the analysis.
For both cancer epidemiology and the study of arsenic metabolism, this is a substantially larger population for which urinary arsenic speciation has been determined at such exposure levels than previously reported. Exposure to lifetime average arsenic concentration in residential drinking water in this population ranged from 0.1 to nearly 200 µg/l, with a mean of 12, with 7% above the earlier water standard of 50 µg/l.
We have found a positive association between exposures to arsenic and all three cancers sites, basal cell carcinoma (BCC) of the skin, bladder cancer, and kidney cancer. The relationship is stronger for skin cancer, for which there were larger numbers and thus more statistical power to detect an effect.
Different exposure metrics show the most significant fit for each cancer site: BCC is most strongly associated with life time average arsenic concentration, bladder cancer with cumulative arsenic dose and kidney cancer with peak arsenic daily dose rate. However, such differences in fit for different exposure indicators may be just a chance finding.
Arsenic species in urine produced by human metabolism of ingested inorganic arsenic were determined in the ASHRAM population. Overall, about 7% of the urinary arsenic was inorganic arsenic, 16% methylarsonic acid (MA), and 77% dimethylarsinic acid (DMA). Compared to previous reports on populations exposed to arsenic via drinking water, the present study shows a lower percentage of inorganic arsenic and a higher percentage of DMA.
This was considered as mainly due to intake of DMA via food. The observed relative proportions of DMA and MA modify the cancer risk attributed to arsenic: cancer risk is increased for the subgroups with low percentage of DMA and/or a high percentage of MA, significantly so in the case of BCC. The observed effect modification is consistent with accepted toxicological models.
Advances were made in the field of arsenic chemistry, with clarification of the probable artefactual nature of MA(III) measurements reported in the recent literature, and the identification of two previously unknown arsenic compounds. Advances were made in the field of cancer genetics, with identification of genetic polymorphisms associations with risk of BCC.
In conclusion we have found a significant association between arsenic and cancer risk and demonstrated the role of inter-individual differences in arsenic metabolism in modifying the cancer risk related to arsenic, a useful finding for risk assessment models.