Science 25 March 2011:
Vol. 331 no. 6024 p. 1504
NEWS & ANALYSIS
Devastation in Japan
The ongoing leaks from Japan's crippled Fukushima Daiichi nuclear power plant have raised concern that some workers and even the public could be exposed to dangerous levels of radiation. So far, officials have said that levels outside the plant are low. But how do they know how much radiation is harmful? Risk calculations are based heavily on a 63-year study of 94,000 people who survived the two atomic bombs dropped on Japan in August 1945. It is one of the largest, longest population studies ever done; for radiation safety, it is the gold standard. Its breadth and precision are “magnificent,” says John Boice, scientific director of the International Epidemiology Institute in Rockville, Maryland, and former chief of radiation epidemiology at the U.S. National Cancer Institute (NCI) in Bethesda, Maryland.
Up to 200,000 people in Nagasaki and Hiroshima died soon after the bomb blasts— some from radiation sickness—but more survived. To understand the delayed effects of radiation, the U.S. National Academy of Sciences launched a joint study with Japan of bomb survivors, using a 1950 census to track them down. Initial case reports of cataracts, leukemia, and birth defects eventually became a long-term study to follow cancer and other illnesses in about 94,000 survivors and 26,000 unexposed residents. It was run by a U.S. and Japanese–funded agency eventually named the Radiation Effects Research Foundation (RERF).
Researchers “spent a huge amount of time reconstructing where people were ATB, at time of blast,” says epidemiologist Richard Monson of Harvard University. They asked subjects whether they were inside or outdoors, near a window, upstairs or downstairs, and which direction they faced. They even constructed a mock Japanese village in the Nevada desert, hoisted a uranium reactor up in a tower, and measured the neutrons it spewed to study the movement of radiation through the air and into buildings. To fine-tune gamma-radiation estimates, researchers tested ceramic roof tiles on Japanese houses for a high-energy electron signature, says Tore Straume of NASA Ames Research Center in Mountain View, California. The dosimetry underwent several revisions over the decades.
Other RERF researchers monitored the health of survivors. Using death records and cancer registries, they soon documented leukemias, particularly in the young, tallying 219 deaths by 2002 in people receiving a significant exposure—a 45% rise above the number expected. It appeared to peak in 1950. By the 1970s, researchers were tracking an elevated rate of solid cancers; “it looks like it persists for a lifetime,” says NCI epidemiologist Kiyohiko Mabuchi.
Still, “people's perceptions of cancer caused are probably different from reality,” says biostatistician Dale Preston of Hirosoft International, who worked at RERF for 23 years. Leukemias were relatively rare, and although by 1998 about 7851 survivors in the study who received significant exposures developed solid cancers, only 850 cases, or 11%, have been attributed to radiation. (The cancer risk was about 50% higher for those who received at least 1 sievert of radiation; the risk drops with dose to 2% below 0.1 sievert. Lifetime risks are lower; for an exposure of 0.250 sieverts, the allowed limit for workers at the Daiichi plant, the increased risk of ever developing cancer is about 2.5%, Boice says.)
Data from hundreds of medical studies have been used to bolster the A-bomb– survivor results. In the early 20th century, before the risks were recognized, radiation was used to diagnose or treat everything from mastitis to tonsillitis—and some patients developed cancer, Boice notes. Studies of workers—such as women who applied radium paint to clock dials and later developed bone cancer—also proved useful. Studies of about 21,000 workers exposed to radiation starting in 1948 at the former Soviet Union's Mayak nuclear weapons plant, and of 30,000 villagers nearby along the Techa River, are proving “important,” Preston says: The cohorts received a wide range of radiation doses, and many workers inhaled plutonium, a long-lived radioisotope absent in the Japanese A-bomb survivors.
These studies and others of nuclear workers “in general have supported estimates from the A-bomb survivors,” says NCI's Ethel Gilbert. Controversy remains, however, about whether the bomb survivors' brief, one-time exposure would be as harmful if spread over many years. “It's the one major unanswered question,” Boice says.
Studies of nuclear accidents have been less useful for estimating dose responses, although they confirm that it's hard to see health effects from low-level exposures. The 1979 accident at Three Mile Island in Pennsylvania exposed the nearby population to a “trivial” amount of radiation, Boice says; health effects were not detected. The 1986 Chernobyl accident, on the other hand, spewed iodine-131 and cesium-137 for 10 days in a plume that reached 5 million people. Researchers expect that 4000 excess cancer deaths will eventually result. But precise dose information is lacking even for the “liquidators,” the 600,000 workers who helped clean up, says Mabuchi, making it difficult to link exposure to disease.
The only clear health effect among the public from Chernobyl so far has been more than 6000 cases of thyroid cancer (15 of them fatal), mainly in people who as children and adolescents drank milk from cows that fed on grass tainted with iodine-131. This should “not be a problem in Japan” because contaminated milk and vegetables are being removed from the food supply, says radiologist Fred Mettler of the University of New Mexico, Albuquerque, a consultant to the United Nations on the Chernobyl disaster.
Studies of the A-bomb survivors (40% of the original group are still living) continue at RERF. A few of its 45 researchers are helping Japanese officials monitor the population near the Daiichi plant, says RERF vice chair and research chief Roy Shore. They're also discussing a possible study of the hundreds of workers involved in keeping the plant under control. If the Fukushima crisis becomes another Chernobyl, Shore says, “we'd certainly want to compile all the data we could.”