Chronic exposure to inorganic arsenic via inhalation or drinking water ingestion has been convincingly linked to cancer of the lung, skin, and urinary bladder (1). This conclusion mainly relies on long-term exposures, and little is known about the consequences of developmental exposures. During the summer of 1955, mass arsenic poisoning of bottle-fed infants occurred in the western part of Japan because of contaminated milk powder (2). Although the milk powder was distributed countrywide, Okayama was the most severely affected prefecture, with a consumption of at least 33 500 one-pound cans of milk powder (3). Although Okayama represents less than 2% of the Japanese population, the 2000 infants who were poisoned or died in Okayama represent 17% of the total number of affected children (3). The dose of poisoned milk powder needed to cause clinical poisoning in an infant corresponds to five cans (2), and the number of possible acute poisonings in Okayama could, therefore, have been three times the official estimate.
We examined the cancer mortality in Okayama by abstracting data for 5-year age and calendar periods from official vital statistics during the period of 1970–2006. We compared the cancer mortality in the 5-year birth cohort that included the subjects born before milk poisoning (exposed group) with a similar cohort born after the milk poisoning (control group). Because the size of a 5-year birth cohort in Okayama is approximately 150 000, the number of clinically poisoned subjects may represent at least 4% of the cohort. After adjustment for the mortality for the country as a whole, we compared the cancer mortality in the exposed cohort with the next (younger) 5-year birth cohort within the same 5-year calendar period and with the same 5-year age group during the subsequent 5-year calendar period. Mortality ratios and their confidence intervals were calculated for total cancer and individual sites.
Both comparisons showed that the mortality from total cancers and lung cancer were not elevated in the exposed cohort, but both skin cancer (including melanoma) and liver cancer were increased (Table 1). Unexpectedly, mortality ratios of pancreatic cancers (all age groups) and leukemia (ages 30–39 years) were elevated in both comparisons. Although little attention has been paid to the pancreas as a possible target of arsenic exposure, arsenic has been reported to induce type 2 diabetes mellitus (4), which may be a risk factor for pancreatic cancer (5). The association with leukemia is consistent with a previous environmental study (6), and the observed narrow latency interval is similar to observations in regard to leukemogenic chemicals (7).
Table 1.
Numbers of deaths and mortality ratios (MRs, with 95% confidence intervals [CIs]) for total cancer and individual sites for the exposed birth cohort from Okayama prefecture, Japan*
| Period or group | Total | Liver | Pancreas | Lung | Skin | Leukemia |
| Same 5-y calendar period | ||||||
| No. of deaths in exposed group at age 30–34 y | 114 | 3 | 5 | 4 | NA | 19 |
| MR during 1985–1989 compared with age 25–29y† (95% CI) | 1.06 (0.89 to 1.28) | 0.26 (0.08 to 0.81) | NE | 0.41 (0.15 to 1.09) | — | 1.67 (1.07 to 2.63)‡ |
| No. of deaths in exposed group at age 35–39 y | 158 | 12 | 8 | 9 | NA | 9 |
| MR during 1990–1994 compared with age 30–34 y† (95% CI) | 0.90 (0.77 to 1.05) | 0.55 (0.31 to 0.96) | 2.26 (1.13 to 4.51)‡ | 0.28 (0.15 to 0.54) | — | 1.47 (0.77 to 2.83) |
| No. of deaths in exposed group at age 40–44 y | 271 | 26 | 13 | 31 | 4 | 9 |
| MR during 1995–1999 compared with 35–39 y† (95% CI) | 1.04 (0.92 to 1.17) | 0.78 (0.53 to 1.14) | 1.04 (0.61 to 1.80) | 0.90 (0.63 to 1.28) | NE | 1.21 (0.63 to 2.33) |
| No. of deaths in exposed group at age 45–49 y | 438 | 50 | 30 | 49 | 3 | 12 |
| MR during 2000–2004 compared with 40–44 y† (95% CI) | 0.93 (0.85 to 1.02) | 2.49 (1.89 to 3.29)‡ | 1.47 (1.02 to 2.10)‡ | 0.64 (0.48 to 0.85) | NE | 0.55 (0.31 to 0.97) |
| No. of deaths in exposed group at ages 15–54 y | 1408 | 125 | 86 | 130 | 10 | 86 |
| Total MR during 1970–2006 compared with 10–49 y† (95% CI) | 0.94 (0.89 to 0.99) | 0.95 (0.80 to 1.14) | 1.52 (1.23 to 1.87)‡ | 0.60 (0.50 to 0.71) | 4.40 (2.37 to 8.18)‡ | 0.82 (0.66 to 1.01) |
| Same 5-y age group | ||||||
| No. of deaths in exposed group at age 30–34 y | 114 | 3 | 5 | 4 | NA | 19 |
| MR comparing 1985–1989 with 1990–1994† (95% CI) | 1.12 (0.93 to 1.34) | 0.32 (0.10 to 0.99) | 3.26 (1.36 to 7.84)‡ | 0.35 (0.13 to 0.92) | — | 3.46 (2.21 to 5.43)‡ |
| No. of deaths in exposed group at age 35–39 y | 158 | 12 | 8 | 9 | NA | 9 |
| MR comparing 1990–1994 with 1995–1999† (95% CI) | 1.06 (0.91 to 1.24) | 0.78 (0.44 to 1.38) | 1.51 (0.75 to 3.02) | 0.58 (0.30 to 1.12) | — | 1.29 (0.67 to 2.48) |
| No. of deaths in exposed group at age 40–44 y | 271 | 26 | 13 | 31 | 4 | 9 |
| MR comparing 1995–1999 with 2000–2004† (95% CI) | 0.95 (0.84 to 1.07) | 2.39 (1.63 to 3.51)‡ | 1.37 (0.79 to 2.35) | 0.71 (0.50 to 1.00) | NE | 0.48 (0.25 to 0.92) |
| No. of deaths in exposed group at age 15–54 y | 1115 | 92 | 56 | 96 | 7 | 80 |
| Total MR comparing 1970–2001 with 1975–2006† (95% CI) | 0.96 (0.90 to 1.02) | 0.97 (0.79 to 1.19) | 1.41 (1.08 to 1.83)‡ | 0.64 (0.53 to 0.79) | 2.19 (1.04 to 4.59)‡ | 0.81 (0.65 to 1.01) |
NA = data before 1994 are not available; NE = not estimated due to small numbers.
Total mortality is for 1970–2006, but individual 5-year comparisons are only shown for 1985–2004 because of small numbers of deaths for ages 15–29 years.
Statistically significantly elevated (P < .05).
Serious dilution is present in these ecological data on survivors of the poisoning incident, but our findings suggest that pancreatic cancer and leukemia could be linked to high-level exposures to inorganic arsenic during infancy. This observation therefore extends previous findings of increased rates for skin, lung, liver, and urinary tract cancer in populations with long-term arsenic exposures at lower levels.
Footnotes
We thank Katherine Herz, Saori Kashima, and Michiyo Yamakawa who helped us to conduct the study and prepare for the manuscript.
References
- 1.Grosse Y, Baan R, Straif K, et al. A review of human carcinogens-Part A: pharmaceuticals. Lancet Oncol. 2009;10(1):13–14. doi: 10.1016/s1470-2045(08)70286-9. [DOI] [PubMed] [Google Scholar]
- 2.Dakeishi M, Murata K, Grandjean P. Long-term consequences of arsenic poisoning during infancy due to contaminated milk powder. Environ Health. 2006;5:31. doi: 10.1186/1476-069X-5-31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hamamoto E. Report of Arsenic Poisoning Incident Due to Powdered Milk in Okayama Prefecture [in Japanese] Okayama, Japan: Department of health, Okayama prefecture, Okayama; 1957. [Google Scholar]
- 4.Diaz-Villasenor A, Burns AL, Hiriart M, Cebrian ME, Ostrosky-Wegman P. Arsenic-induced alteration in the expression of genes related to type 2 diabetes mellitus. Toxicol Appl Pharmacol. 2007;225(2):123–133. doi: 10.1016/j.taap.2007.08.019. [DOI] [PubMed] [Google Scholar]
- 5.Lowenfels AB, Maisonneuve P. Epidemiology and risk factors for pancreatic cancer. Best Pract Res Clin Gastroenterol. 2006;20(2):197–209. doi: 10.1016/j.bpg.2005.10.001. [DOI] [PubMed] [Google Scholar]
- 6.Hinwood AL, Jolley DJ, Sim MR. Cancer incidence and high environmental arsenic concentrations in rural populations: results of an ecological study. Int J Environ Health Res. 1999;9(2):131–141. [Google Scholar]
- 7.Linet MS, Devesa SA, Morgan GJ. The leukemias. In: Schottenfeld D, Fraumeni JF, editors. Cancer Epidemiology and Prevention. 3rd ed. Oxford, New York: Oxford University Press; 2006. pp. 841–871. [Google Scholar]