Results
Incidence of all cancers. Incidence for total cancers for children < 5 yr during 1988 to
1997 was higher than the SEER rate near all 14 nuclear plants in our study (Table 4).
The rate for all 49 counties combined was 22.51 per 100,000, or 11.4% greater than the
SEER rate (p < 0.0002). The smallest excess was near the Salem/Hope Creek complex
(+0.7%); the largest occurred near both the Turkey Point and St. Lucie facilities in
Florida (+29.1%).
Cancer incidence in children 5-9 yr for 1988 to 1997 exceeded the SEER rate for 13 of
the 14 areas. The rate for the study counties was 12.15 per 100,000--12.5% higher than
the SEER rate of 10.80 (p < 0.002). The smallest excess was found near the Millstone
reactors in Connecticut (+2.2%), and the largest occurred near St. Lucie (+73.6%).
Incidence near the Crystal River facility in Florida was 6.5% below the SEER rate.
Combining the age groups yields an incidence rate of 17.42 per 100,000--12.4% above
the SEER rate (p < 0.00001). The excess incidence near 3 of the plants (Oyster Creek,
St. Lucie, and Turkey Point) was statistically significant; near the Indian Point and
Brookhaven facilities it reached borderline significance (p < 0.08 and p < 0.07,
respectively). Although county-specific totals are not shown, considerable variation in
rates exists, in part because of the relatively small numbers of cases involved. Still, the
incidence rate for those 0-9 yr of age exceeded the U.S. rate in 38 of the 49 study
counties.
Childhood cancer incidence < 30 mi (48 km) from nuclear reactors was compared with
rates for the remaining counties in the states in which reactors are located. Several
adjoining, less-populated states (New Jersey and Delaware, Connecticut and Rhode
Island, Massachusetts and New Hampshire) were combined to ensure adequate
statistical power. For each of 6 states and combinations of states, cancer incidence for
those 0-9 yr in the counties near reactors was higher than in other counties in the state
(Table 5). The total excess incidence derived from comparing the counties near reactors
with those in the rest of the state, or state combinations, was 5.0% (p < 0.04).
Elevated rates for the New York and Pennsylvania nuclear counties are of borderline significance
(p < 0.055 and p < 0.07, respectively).
Total cancer incidence by race. U.S. black and Hispanic children < 20 yr of age have
cancer incidence rates 23% and 10% below that for whites, respectively. (44) To assess
the effect of race on childhood cancer incidence near nuclear plants, incidence data from
Pennsylvania counties near nuclear plants were studied (the Pennsylvania registry
makes county statistics for whites and blacks more readily available than do registries in
other states). Using SEER data, the 1988 to 1997 U.S. cancer incidence rates for white
and black children 0-9 yr were calculated at 15.88 and 13.28 per 100,000, respectively.
For the 23 Pennsylvania counties located close to reactors, childhood cancer rates
exceeded U.S. rates for both whites and blacks (Table 6).
Incidence of leukemia. We examined the incidence of childhood leukemia in the 23
counties near 5 nuclear plants in Pennsylvania (Table 7). These regions account for
slightly more than half the state's population. Leukemia incidence in the state's nuclear
counties exceeded the U.S. rate by 10.8%; the rate for the remainder of the state was
11.5% below the U.S. rate (p < 0.01). For all other cancers, virtually no difference was
seen between nuclear and non-nuclear counties, even though both exceeded the
national rate (by 2.6% and 3.2%, respectively).
[...]
This study found a consistent pattern of increased childhood cancer incidence in all
study areas < 30 mi (48 km) from nuclear plants in the eastern United States. Our
findings support the biologically plausible concept that susceptibility to carcinogens, such
as radioactivity, is greatest in utero and in early childhood. They also support numerous
analyses documenting elevated childhood cancer rates near nuclear facilities in the
United States and other nations. The finding that cancer incidence for children < 10 yr is
12.4% greater in the study counties than the U.S. as a whole suggests that emissions
from nuclear power plants may be linked with 1 of 9 local cases of childhood cancer.
These descriptive epidemiological findings suggest a relationship between radioactive
nuclides and childhood cancer and should be taken seriously in future research.
