STRONTIUM-90 IN BABY TEETH AS A FACTOR IN EARLY CHILDHOOD CANCER

Jay M. Gould, Ernest J. Sternglass, Janette D. Sherman,
Jerry Brown, William McDonnell, Joseph J. Mangano

International Journal of Health Services
Volume 30, Number 3, Pages 515-539, 2000
Copyright Baywood Publishing Co., Inc

Methods
In 1996, the Radiation and Public Health Project (RPHP), a New York City-based research organization, began collecting baby teeth and analyzing them for Sr-90 concentrations. The study proposes to document present levels and historical trends of Sr-90 in human bodies since the end of all atmospheric bomb tests in 1980. Thus, RPHP will be able to compare current Sr-90 patterns (when most of this radionuclide is taken up from nuclear power reactor emissions) with those of 35 to 50 years ago (when most in vivo Sr-90 represented bomb test fallout).

The initial geographic focus of the study was Suffolk County, New York, a 922-square mile area making up the eastern three quarters of Long Island with a 1990 population of 1,321,864 (See Figure 3).

Suffolk was selected because for more than a decade, the county has been the subject of numerous studies in the medical literature, due to consistently high levels of breast cancer incidence and mortality. (20) (21) Rates in the suburban areas of Suffolk and Nassau County (adjacent to the western border of Suffolk) began to exceed even the high rate in New York City beginning in the late 1970s. Moreover, a recent study by RPHP researchers had linked the rise in breast cancer in both Suffolk and Nassau to fission products in the diet and water produced by airborne releases from nearby nuclear plants. (22)

Suffolk is the site of the Brookhaven National Laboratory (BNL), a research facility located near the town of Upton in western Suffolk some sixty miles from the eastern end of Long Island. BNL has operated two or three nuclear research reactors at various times since 1951, which have released significant quantities of radioactive elements into the local air, ground and surface waters.

As shown in Figure 3, Suffolk is also proximate to several other reactors operating in the 1980s and 1990s. The earliest to begin operation was the Indian Point Plant Unit 1 that started in 1962, followed by the much larger Units 2 and 3 in 1973 and 1976. Located on the Hudson River some 30 miles north of Manhattan, the prevailing winds from the northwest in the winter months move the airborne releases from Indian Point towards western Suffolk, as indicated by the wind distribution measured at BNL shown in Figure 3.

Another nuclear plant (Oyster Creek) is located some 60 miles to the southwest on the New Jersey coast near the town of Toms River that began operation in 1969. The prevailing winds in the summer months move from the southwest towards western Long Island, so that the plumes of airborne activity from Oyster Creek cross those from Indian Point in the area of Suffolk just to the west of BNL. Thus the western part of Suffolk receives airborne emissions from three nuclear sites for most of the year, making it the most heavily exposed part of the county. Furthermore, the Oyster Creek plant reported the second highest airborne releases of radionuclides with a physical half life of eight days or more, out of 76 nuclear sites in the nation, with a total of 76.8 curies since 1970. This total far exceeds the 14.2 curies reported for the Three Mile Island Unit 2 during the 1979 accident. (23)

It is in this highly exposed area of western Suffolk that 17 out of 18 communities with breast cancer incidence rates above the county average are located, as reported by the New York State Department of Health for the period 1978-87. (22) It is also the area where 19 recent cases of an extremely rare form of childhood cancer (rhabdomyosarcoma) have been identified. (24) Finally, in 11 contiguous five-digit ZIP code areas, which had a 1990 population of 162,187 and was located 10-15 miles northwest of BNL, 79 hospital discharges for cancer in residents age 0-9 occurred in 1997. In comparison, 0 admissions were from easternmost Suffolk, known as the North and South forks (1990 population = 65,416). (25)

Two nuclear plants in Connecticut to the north of Suffolk are Haddam Neck near Middletown and Millstone near New London. The three Millstone reactors are only 11 miles north of the northeastern tip of Suffolk County. Although the prevailing winds are directed towards the eastern part of Suffolk only in the winter months, these reactors, like Oyster Creek, have had a substandard record of radioactive emissions, releasing 32.6 curies of long-lived fission products since 1970 - the third highest release among all U.S. nuclear plants. (23) In the winter of 1995-96, the Nuclear Regulatory Commission ordered all three Millstone reactors closed for safety violations. Unit 1 was shut down permanently, while the other two did not operate for 2 ½ and 3 ½ years while extensive physical and management improvements were made. Thyroid cancer, the risk of which is known to be increased after exposure to radioactive iodine found in all reactor emissions, reached significantly high levels in New London (near Millstone) in the late 1980s and early 1990s before the plant was closed. (26)

One other nuclear plant along the northeastern Atlantic coast that is less than 100 miles from the Eastern end of Suffolk is the Pilgrim plant 25 miles south of Boston near Plymouth MA. It began operation in 1972 and experienced a series of severe problems, including a major accidental release of radioactive gases in 1982 that was detected all over New England. (27) An abnormally high incidence of leukemia was reported around this plant. (28)

Precipitation brings down 90 percent of airborne radioactivity to earth, introducing it into the food chain. (29) Thus the rain and snow in Suffolk recharges the aquifers supplying the drinking water for the wells all over Suffolk county with contamination from radioactive particles released from these facilities.

Deciduous teeth, usually shed between the ages of 5 and 12, were selected because they represent a stable calcified tissue similar to bone but in which rates of turnover, exchange, remodeling and accretion are minimal or absent. Thus the deciduous tooth may be considered a stable structure and can thus be taken to reflect the mineral composition acquired at the time of its formation. (5) (6) It can therefore provide information concerning the amount of Sr-90 uptake during the year of the child's birth. A portion of the calcium and Sr-90 is drawn from the mother's bone in addition to the mother's dietary intake during pregnancy, and another portion is derived from the infant's diet in the year after birth for all types of teeth. (5)

Thus, for all the purposes of the present study, there is no need to measure incisors, cuspids, and molars separately. Wide variations in Sr-90 occur due to dietary differences and the time of the year when the child is born, (5) so that many teeth are needed to obtain meaningful yearly average values. Given an adequate number of teeth for a given area (i.e., three digit ZIP code area), it is possible to obtain accurate historical data not only on the body burden of Sr-90 for the child but also for the mother in the year of the infant's birth. Thus, deciduous teeth provide a much more accurate indication that individuals in certain locations received radiation exposures from fission products at a given time. This information cannot be determined from single monthly or yearly milk measurements aggregated for a large geographic area.

Following previous studies, the RPHP effort focused on Sr-90 because of the reliability of detecting this long-lived radionuclide, due to its extended half-life of 28.7 years. Relative levels of this pure beta particle-emitting isotope are also a proxy for several dozen other short and long-lived radioactive fission products emitting gamma rays, and alpha and beta particles, found in bomb test fallout and power reactor emissions. Discarded baby teeth are used because of the greater ease of collecting large numbers rather than from bones from autopsies. The majority of the teeth being studied in the initial phase are a result of a mass mailing to 15,000 randomly selected households with children age 6 to 18 in early 1999. Donors also submitted teeth after learning of the project from print, radio, and television stories. A toll-free telephone number and a site on the World Wide Web are other ways in which the public has been able to obtain information on tooth donation. *

* The web site is https://radiation.org [this site].

Each potential donor is mailed a package containing a letter and several articles describing the study. An envelope in which the donor can send teeth to RPHP is also enclosed. In addition to the tooth (some contributors send multiple teeth), the following information is requested on the envelope:

- Mother's name
- Phone number
- Address
- Child's name
- Birth date (month, day, year)
- Birth weight (pounds, ounces)
- Location where mother carried the baby (city, state, county, ZIP code)
- Location where child was born (city, state, county, ZIP code)
- Location during first year of life (city, state, county, ZIP code) - also second and third year
- Water source (well, municipal water, bottled water, other)
- Mother's age at birth of child
- Age of child when tooth was lost/date of tooth loss
RPHP assures donors of the confidentiality of all information identifying mothers and children.

The initial mailing of envelopes targeted selected populations, representing those exposed to different degrees of radioactivity from reactors in and around Long Island according to their geographic location. The more heavily exposed populations include children living in western Suffolk County NY (ZIP codes beginning 117), mid-New Jersey, near the Oyster Creek reactor (ZIP codes beginning 087), and the greater Miami area, near the two Turkey Point reactors (ZIP codes beginning 330 and 331). Less exposed populations were residents of parts of Eastern Suffolk (ZIP codes 119) and parts of Queens, New York and northwestern New Jersey. As of October 1, 1999, about 1400 teeth had been received by RPHP, a majority of them a direct result of the early 1999 mailing. This response rate for this unsolicited mailing has been greater than three percent.

Upon receipt of envelopes containing teeth, RPHP staff assigns a unique control number to each tooth and logs it into a computerized data base. Teeth are periodically sent in batches to a radiochemistry laboratory in Waterloo, Ontario, Canada. Laboratory personnel document Sr-90 concentrations by separately measuring Sr-90 activity (in picocuries) and calcium weight (in grams) in the teeth. Waterloo researchers are blinded from any information about each tooth. The techniques employed for this purpose are explained in Appendix 1.

Sr-90 readings for each tooth at the time it was lost by the child are converted to levels at birth, creating a standardized measure that can be compared regardless of how old the tooth is. This conversion is made using the decay rate of a 28.7 year half-life for the radionuclide, using the month of birth and the month of tooth analysis.

Results
As of October 1, 1999, a total of 515 teeth had been analyzed for Sr-90 concentration. Of this number, 476 were from children born during the years 1979 to 1994, but with only 15 teeth in 1993-94, not enough to provide high statistical reliability for these years.

Table 1 indicates the geographic distribution of the teeth, based on where the mother carried the child, along with the average and maximum concentrations for each of four ZIP code areas accounting for 374 of the most recent teeth analyzed. Few teeth from less-exposed control areas in New York and New Jersey have been received, and are thus omitted from this analysis.

The mean concentration for all 476 teeth is 1.50 pCi Sr-90/g Ca, roughly equivalent to that found in St. Louis children born in 1956, five years after atmospheric testing in Nevada had commenced. The standard error for this average is +/- 0.03. While the mean level for Dade County FL (Miami) teeth is highest (2.80), averages for the entire period 1979-94 from all four areas greatly exceed the level of about 0.2 pCi Sr-90/g Ca. This level would be expected if post-Test Ban Treaty (1964-70) declines observed in St. Louis teeth had continued (Figure 4). Only 135 (28.4%) of the teeth had 0.2 pCi Sr-90/g Ca, while 67 (14.1%) of the sample had readings of 3.0 or higher, about 15 times greater than the projected values. The largest single value was 17.87 pCi Sr-90/g Ca.

Sr-90 in 476 RPHP Baby Teeth Vs. Projected Levels from St. Louis Teeth 1979-94

Annual averages in RPHP Sr-90 concentrations are variable, but both averages and peak values rose to their highest levels in 1988, two years after the arrival of Chernobyl fallout. Average measurements rose from 1.48 to 1.51, or 2.2% from 1981-87 (n = 190) to 1988-94 (n=276).

Figure 4 displays the RPHP results since 1979, compared to prior measures of Sr-90 in St. Louis baby teeth and in New York City adult dietary uptake. Since both Sr-90 levels in the adult uptake and in the St. Louis baby teeth declined on average by 15.7 percent each year in the years 1964-70, this annual percentage decline is used to project the expected Sr-90 levels in baby teeth from 1970 to 1994. It is clear that the RPHP levels are far higher - in fact about three times higher - than the levels expected based on the assumption that no additional Sr-90 had been introduced into the atmosphere since 1970, a significant difference (p value is infinitesimal). The annual adult dietary uptake in any given year is closely related to the dietary uptake of pregnant women, and thus to the level observed in the baby teeth of children born in that year. Note for example that the adult uptake rose 6 percent (7.8 to 8.3 pCi Sr-90/g Ca) from 1978 to 1979, the year of the Three Mile Island (TMI) accident, and then declined 23 percent, to 5.6 in the following two years when the two TMI reactors were shut down. This accords with the relatively high values found in the RPHP averages in the early 1980s, followed by a rise in 1986 and subsequent decline associated with the arrival and subsequent decay of Chernobyl radiation. Further corroboration is reflected in the observed EPA measures of high radioactivity in milk and water in May and June of 1986. (15)

In-depth analysis can be performed on Suffolk County children, who contributed 304 (64%) of the 476 teeth tested to date. A comparison of the mean annual Sr-90 concentrations in teeth with cancer incidence age 0-4 in the county (30) is made in Figure 5.

Average Sr-90 Concentrations in Baby Teeth, 1980-89 vs. Cancer Incidence Age 0-4, 1983-92 Suffolk County, Three Year Moving Average

A three-year latency between relative radiation level and cancer diagnosis is assumed, i.e., Sr-90 concentrations for the period 1980-89 are matched with cancer incidence for 1983-92. A three year delay between the Suffolk Sr-90 measures at birth and the incidence of cancer 0 to 4 was found to give the best fit. This agrees with the observation that childhood leukemia and brain cancer that make up a major share of all childhood malignancies are most frequently diagnosed between the second and fifth year of life. Three-year averages are used; for example, the cancer rate for 1989 represents cancer cases per population for the years 1988-90 combined. Using three year totals enhances the significance of the analysis, as some individual years may have as few as five teeth, so that annual Sr-90 averages are subject to statistical variation, which is reduced by the three year moving average without affecting turning points. The number of newly-diagnosed cancer cases age 0-4 over three years varies from a low of 41 to a high of 73.

The analysis of Sr-90 in teeth was not conducted for births after 1990. As mentioned, the only birth years with substantial numbers of teeth received are 1991 and 1992. Furthermore, most of the recent teeth being donated are from children born in the 1990s, and this pattern will likely continue into the future. While patterns for 1980s births appear to be set, those for the 1990s may well change in the future, making reporting present totals premature.

The use of three year moving averages for both Sr-90 and cancer incidence for the relatively small number of cancer cases and teeth in recent years, contrasts sharply with the case of the period of major nuclear tests shown in Figure 1. During the peak fallout years of large scale hydrogen bomb tests in 1961 and 1962, the long-lived Sr-90 was injected into the stratosphere, as in the case of the 1961 50 megaton Soviet test, which drifted down into the lower atmosphere over a period of years. This led to a delayed peak of Sr-90 of some two to four years after the immediate exposure of the fetus and young infant to the short-lived fallout brought down in 1961-62. Since childhood cancer for the 0-4 year group is also typically diagnosed some two to four years after birth, the peaks in Sr-90 and 0-4 cancer incidence occurred at about the same time.

By contrast, the releases from nuclear reactors do not reach the stratosphere and instead are brought down by rain or snow in a matter of hours, days or weeks. Sr-90 enters the air that is inhaled and the food chain rapidly, together with the many short-lived isotopes that produce the greatest exposure to the embryo and fetus, and also quickly reaches the drinking water from shallow wells in Suffolk. For nuclear plant releases after the last Chinese atmospheric test of 1980, we would expect a two to four year lag in 0-4 year cancer incidence after Sr-90 is deposited in the newly formed tooth enamel. As a result, Sr-90 in baby teeth can serve as an indicator of the time of principal exposure only after the end of large-scale hydrogen bomb tests. But it is not the principal source of radiation exposure leading to early childhood cancer, which is mainly due to the much more abundant short-lived isotopes transferred to the fetus through the placenta, such as Barium-140, Iodine-131, and Strontium-89 that accompany Sr-90.

Figure 5 provides data on average Sr-90 concentrations in teeth for births 1980-89, and correlates them with cancer incidence age 0-4 for 1983-92. Too few teeth are available for persons born after 1990.

Sr-90 buildup in Suffolk baby teeth during the early 1980s and modest reduction in the late 1980s appear to be matched (three years later) by similar patterns in cancer incidence. The peak levels of Sr-90 during the late 1980s, up about 50% from early in the decade, strongly suggest the important role played by the Chernobyl radiation cloud that reached the U.S. in May 1986. For the entire 10-year period, the two measures are correlated (r = .85, P<.001) (Figure 5). This linkage is similar to that between St. Louis baby teeth and Connecticut childhood cancer in the 1950s and 1960s, discussed earlier.