An unexpected rise in strontium-90 in US deciduous teeth in the 1990s

The Science of the Total Environment 317 (2003) 37–51
0048-9697/03/$ - see front matter
2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0048-9697(03)00439-X

Joseph J. Manganoa,*, Jay M. Gouldb,1, Ernest J. Sternglassc,2, Janette D. Shermand,3,
William McDonnelle,4
aRadiation and Public Health Project, 786 Carroll Street, 9, Brooklyn, NY 11215, USA
bRadiation and Public Health Project, 302 West 86th Street, 11B, New York, NY 10024, USA
cRadiation and Public Health Project, 4601 Fifth Avenue 824, Pittsburgh, PA 15213, USA
dP.O. Box 4605, Alexandria, VA 22303, USA
eRadiation and Public Health Project, P.O. Box 60, Unionville, NY 10988, USA
Received 3 March 2003; received in revised form 14 March 2003; accepted 11 July 2003
Abstract
For several decades, the United States has been without an ongoing program measuring levels of fission products
in the body. Strontium-90 (Sr-90) concentrations in 2089 deciduous (baby) teeth, mostly from persons living near
nuclear power reactors, reveal that average levels rose 48.5% for persons born in the late 1990s compared to those
born in the late 1980s. This trend represents the first sustained increase since the early 1960s, before atmospheric
weapons tests were banned. The trend was consistent for each of the five states for which at least 130 teeth are
available. The highest averages were found in southeastern Pennsylvania, and the lowest in California (San Francisco
and Sacramento), neither of which is near an operating nuclear reactor. In each state studied, the average Sr-90
concentration is highest in counties situated closest to nuclear reactors. It is likely that, 40 years after large-scale

atmospheric atomic bomb tests ended, much of the current in-body radioactivity represents nuclear reactor emissions.

2003 Elsevier B.V. All rights reserved.
Keywords: Radiation; Strontium-90; Nuclear reactors; Deciduous teeth (baby teeth)
*Corresponding author: Tel.: q1-718-857-9825; fax: q1-718-857-4986.
E-mail addresses: odiejoe@aol.com (J.J. Mangano), jaymgould@aol.com (J.M. Gould), mssejs@aol.com (E.J. Sternglass),
toxdoc.js@verizon.net (J.D. Sherman), bill@oldbooks.net (W. McDonnell).
1 Tel.: q1-212-496-6787; fax: q1-212-362-0348.
2 Tel.yfax: q1-412-681-6251.
3 Tel.: q1-703-329-8223; fax: q1-703-960-0396.
4 Tel.: q1-845-726-3355.
38 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
Introduction
Since man-made fission products were first
released into the environment in the mid-1940s,
determining in vivo levels of these radioisotopes
has challenged scientists. Hundreds of radioisotopes
are created in nuclear weapon detonations
and in nuclear reactor emissions. Many of these
are short-lived, and therefore highly unlikely to
track in vivo. Collecting samples of longer-lived
isotopes often involves invasive processes such as
autopsies and biopsies, making collection of significant
samples time-consuming and costly.
In the US, whose government conducted 206
atmospheric tests of nuclear weapons from 1946
to 1962 (100 in Nevada, 106 in the South Pacific)
(Norris and Cochran, 1994), the federal government
instituted programs measuring strontium-90
(Sr-90) concentrations in vertebrae. One focused
on deceased adults (begun 1954, 3 cities, ;50
bones per year) (Klusek, 1984), while the other
included deceased children and adolescents (begun
1962, 30 cities, ;300 bones per year) (Baratta et
al., 1970). Both showed increases to a peak in
1964, just after the Partial Test Ban Treaty was
signed, and a dramatic decline in the mid- and late
1960s.
The largest-scale US program studying in-body
radioactivity was conducted in St. Louis. Kalckar
suggested that large numbers of deciduous teeth
could be collected and tested to examine the
buildup of fallout from bomb tests (Kalckar,
1958). A coalition of St. Louis medicalyscientific
professionals and citizens collected over 300 000
teeth from local children from 1958 to 1970.
Results from St. Louis were similar to the two
bone programs, i.e.
– A 55-fold rise in average millibecquerels (mBq)
of Sr-90 per gram calcium at birth (7.4–408.1)
took place for 1949–1950 births (before largescale
tests began) to 1964 births ( just after the
largest-scale bomb tests ended).
– A 50% decline in Sr-90 concentrations in St.
Louis fetal mandibles occurred from 1964 to
1969 births. This far exceeded the expected 9%
reduction suggested by the 28.7 year half-life
of Sr-90 (Rosenthal, 1969).
After the bone and tooth studies showed such a
rapid post-1964 decline, federal funding was terminated
for each program, and work ceased. The
tooth study ended in 1970, the child bone study in
1971 and the adult bone study in 1982.
The US studies were accompanied by similar
international efforts. Each independently confirmed
the American findings of rapid increases in
teeth until 1964, including studies in Czechoslovakia,
Denmark, Finland and Scotland (Santholzer
and Knaifl, 1966; Aarkrog, 1968; Rytomaa, 1972;
Fracassini, 2002). Another study in Finland duplicated
the rapid post-1964 plunge in Sr-90 (Kohlehmainen
and Rytomaa, 1975). No nation
maintained an ongoing program, but after the
Chernobyl accident, reports from Germany, the
Ukraine and Greece documented a substantial rise
in Sr-90 in baby teeth after the April 1986 disaster
(Scholz, 1996; Kulev et al., 1994; Stamoulis et
al., 1999). Another study examined Sr-90 in teeth
from children who lived proximate to the Sellafield
nuclear installation in northwestern England;
results are addressed in Section 4 (O’Donnell et
al., 1997).
With no program of in vivo radioactivity to
gauge the burden on the body, levels in the
environment can be used as a proxy measure. In
the past, patterns of Sr-90 in baby teeth were
roughly equivalent to those of Sr-90 in milk
(Rosenthal et al., 1964). The US government
(beginning 1957) began publicly reporting monthly
levels of a variety of radionuclides in milk and
water in 40–60 US locations. However, a number
of these radioisotopes, including Sr-90, strontium-
89, cesium-137, barium-140 and iodine-131 were
discontinued in the early 1990s (National Air and
Radiation Environmental Laboratory, 1975–2001).
One measure that is still publicly reported is the
concentration of gross beta particles in precipitation.
A reduction in average beta levels reversed
after 1986–1989. While the most recent 4-year
period still features incomplete data, thus far the
increase from 1986–1989 to 1998–2001 has been
53.8%. This difference is significant at P-0.0001,
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 39
Table 1
Trend in gross beta in precipitation in average millibecquerels per liter of water in 60 US cities, 1978–2001
4-year period Months Number of Average betaa Percent change,
available measurements 1986–1989 to 1998–2001b
1978–1981 36 640 211
1982–1985 48 1299 63
1986–1989 46c 1845 58
1990–1993 48 1892 59
1994–1997 48 1696 63
1998–2001 27 836 89 q53.8% (P-0.0001)
The P value indicates that the chance that the increase is due to random chance is fewer than 1 in 10 000. Source: Environmental
Protection Agency, Environmental Radiation Data, quarterly volumes.
a Average millibecquerels of gross beta per liter of precipitation (reported by EPA as picocuries; to convert to millibecquerels,
multiply by 37). Before 1996, figures were reported as nanocuries per meter squared at a particular depth (in millimeters); to convert
to pCiyl, multiply nCi per meter squared times 1000, then divide by millimeters; then multiply by 37 to obtain millibecquerels.
b Calculation of change beginning with lowest average (1986–1989) to most current.
c Excludes May and June 1986, heavily affected by short-lived Chernobyl fallout.
i.e. the probability of this increase due to random
chance is less than 1 in 10 000 (Table 1).
The lack of an ongoing program measuring in
vivo radioactivity levels and an unexpected, sustained
rise in environmental beta concentrations
warrant a resumption of testing in vivo Sr-90 and
perhaps other radioisotopes, first instituted during
the era of atmospheric nuclear weapons testing.
In 1996, the Radiation and Public Health Project
(RPHP) began a study of Sr-90 levels in deciduous
teeth, focused on persons living near nuclear reactors.
The goal of this project was to build a current
database of in vivo radioactivity documenting Sr-
90 patterns and trends. While Sr-90 is just one of
hundreds of radioisotopes from fission, it can be
used as a proxy for all fission products, especially
those with extended half-lives.
2. Materials and methods
Earlier reports addressed methods used and initial
findings from the baby tooth study (Gould et
al., 2000a,b; Mangano et al., 2000). These teeth
were processed using a scintillation counter from
the University of Waterloo in Ontario, Canada. In
June of 2000, RPHP leased a Perkin-Elmer 1220-
003 Quantulus Ultra Low-Level Liquid Scintillation
Spectrometer. Introduced in 1995, only
approximately 15–20 units are now in use in the
US (Laxton, Mark, Perkin-Elmer Life Sciences
Inc., 549 Albany Street, Boston MA 02118. Personal
correspondence, May 9, 2002).
The new counter is located on the premises of
REMS, Inc., a radiochemistry laboratory in Waterloo,
and not at the University of Waterloo, thus
changing the level of background radiation. Also,
the method of removing organic material from the
teeth was changed by treating them with hydrogen
peroxide prior to grinding them into powder. This
procedure proved to be more effective in allowing
light produced in the liquid scintillation fluid by
the beta particles emitted by the Sr-90 and its
daughter product, Yttrium-90, to reach the photomultipliers.
This greater efficiency is caused partly
by shifting the spectrum of the light emitted by
the scintillation fluid. As a result of these changes
(the counter, its location, level of background
radiation and method of cleaning teeth), the efficiency
of detecting the very low radioactivity in
single teeth was more than doubled overall. However,
the data lack a consistent factor that could
be used to analyze teeth from both counters together.
Thus, this report will be based solely on the
2089 deciduous teeth tested after June 2000.
RPHP sends teeth to REMS for testing, and Sr-
90 levels are measured individually. Lab personnel
are blinded about all information concerning each
tooth, that is, they know nothing about character40
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
Table 2
Average millibecquerels of Sr-90 per gram calcium (at birth) in deciduous teeth from St. Louis, 1954 and 1959 births (test for
internal consistency)
Batch Average % 1959 Counting error 95% confidence
Sr-90a over 1954 interval
1 1954 61 "10 41–81
1959 121 q98 "13 95–147
2 1954 65 "11 43–87
1959 124 q90 "14 96–152
a Average millibecquerels of Sr-90 per gram of calcium.
istics of the tooth donor. This blinding helps assure
objectivity in results. The laboratory measures the
concentration of Sr-90 by calculating the current
activity (in mBq) of Sr-90 per gram of calcium in
each tooth (mBq Sr-90yg Ca). (See Appendix A
for more specific technical procedures.) The strontium-
to-calcium ratio has been used in the St.
Louis study in the 1960s, and all other recent baby
tooth studies mentioned earlier.
The laboratory returns results to RPHP staff,
who converts the ratio to that at birth, using the
Sr-90 half-life of 28.7 years. The Sr-90yCa ratio
for a single tooth is not a precise number because
a typical baby tooth is small in mass. The counting
error for each tooth is plus or minus 26 mBq, and
somewhat less for the larger teeth.
RPHP conducted several tests to assure the interlaboratory
reliability and internal consistency of
its results. It selected 10 teeth from persons born
in 1954 in St. Louis that were tested both by
REMS and the University of Georgia Center for
Applied Isotope Studies, which operates three
counters of the same model. REMS dried the 10
teeth and ground them into a powder. After testing
for Sr-90 levels, the entire batch was sent to the
University of Georgia, which tested a dissolved
solution of teeth. Both labs were blinded from
each other’s results. The data were relatively comparable.
REMS’ average was 65 mBq Sr-90yg Ca
(CIs43–87), while University of Georgia’s tally
was 79 mBqyg Ca (CIs56–102).
REMS also performed a second test, for internal
consistency. Prior results from the St. Louis study
indicated that average 1959 Sr-90 levels were
considerably higher than those for 1954, due to
buildup in bomb test fallout. RPHP split two
samples of 10 teeth, each into two batches, and
asked REMS to calculate average Sr-90 levels
separately. Results, shown in Table 2, documented
the 1959 average to be 98 and 90% higher than
the 1954 average. Confidence intervals showed
considerable overlap, indicating that study results
are consistent both internally and with the earlier
St. Louis study.
A third test for accuracy involved several dozen
teeth from persons born in the Philippines Islands
1991–1992. This area has never had a nuclear
reactor (for weapons, power or research). It may
have received fallout from Chinese atmospheric
bomb tests, but there were many fewer of these
than US tests. Chinese atmospheric tests ended in
1980, and the last below-ground test occurred in
1993. Thus, Philippino teeth should contain lower
concentrations of this radioisotope than American
teeth.
Thirteen teeth of Philippino children born in
1991 and 1992 were tested. The average concentration
at birth was 75 mBq Sr-90yg Ca, or 41%
lower than the 127 mean level for American
children born in those years.
RPHP collects teeth through voluntary donations,
mostly from parents of children who have
recently shed a deciduous tooth. Donors submit
teeth in envelopes containing identifying information
on the child and parents. RPHP staff assigns
each tooth a unique tracking number. The group
sent nearly 100 000 unsolicited letters appealing
for tooth donation to families with children age
6–17. These mailings occurred in California (Sacramento
and San Luis Obispo counties), Florida
(Dade and Port St. Lucie counties) and New York
(Rockland and Westchester counties). Families
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 41
Table 3
Average millibecquerels of Sr-90 per gram calcium in deciduous teeth (at birth) by state (all persons and persons born after 1979)
State Teeth Average Sr-90a Counting error
All persons
PA 133 155 "14
Oth 492 146 "7
NY 557 141 "6
NJ 271 139 "9
FL 485 131 "6
CA 151 114 "10
TOT 2089 139 "3
Persons born after 1979
PA 130 154 "14
NY 534 138 "6
FL 471 130 "6
Oth 417 130 "6
NJ 244 125 "8
CA 138 108 "10
TOT 1934 132 "3
See Appendix Bfor explanation of error calculation.
a Average millibecquerels of Sr-90 per gram of calcium.
receiving letters were randomly selected by zip
code in each county, that is, every nth family in
each zip code received a letter. Just over 1% of
these mailings were returned with a baby tooth
enclosed.
Teeth are geographically classified by the zip
code where the mother resided during pregnancy,
rather than the current residence. The large majority
of Sr-90 uptake in a baby tooth occurs during
the fetal and early infant periods (Rosenthal,
1969), making zip code during pregnancy the
appropriate geographic identifier.
Other teeth were collected from persons who
became familiar with the project through media
articles and stories, and through the group’s web
site. Thus, the teeth are not necessarily representative
of the US population at large. The vast
majority is concentrated in only five states (California,
Florida, New Jersey, New York and Pennsylvania),
near nuclear reactors. Most were
donated from children who have just recently lost
a tooth, or those between age 5 and 13. Despite
these shortcomings, the large number of teeth will
enable meaningful analysis of average Sr-90 concentrations
to be performed; and any major variations—
by birth year, by state, etc.—will likely be
discernible.
3. Results
3.1. By state
A total of 2089 teeth were tested for Sr-90, and
are discussed in this paper (another 1335 had been
tested previously using a different scintillation
counter and method). As discussed, the two sets
of results are each internally consistent, but not
comparable with each other because of differences
in the counter, its location, level of background
radiation and method of cleaning teeth, so only
the last 2089 teeth are used. Of these, 1592 (77%)
were from children born in the five states mentioned
earlier, each with at least 133 teeth studied.
No other state has more than 34 teeth. Table 3
shows the comparative average Sr-90 concentrations
by state.
Table 3 also displays averages by state only for
persons born after 1979. The large buildup from
above-ground nuclear weapons tests reached a
peak in 1964, and fell by approximately half over
42 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
Table 4
Average millibecquerels of Sr-90 per gram calcium in deciduous teeth (at birth) by proximity to nuclear power plants (persons born
after 1979)
Nuclear power Proximate Average Sr-90a (No. teeth) % Difference
plant, location counties
Proximate Other state
average Sr-90
Indian Point, Buchanan NY Putnam, Rockland, 164 (217) 121 (317) q35.8% P-0.001
(2 reactors, startup 1973, 1976) Westchester, NY "11 "7
Limerick, Pottstown PA Berks, Chester, 168 (98)b 110 (32) q53.2% P-0.03
(2 reactors, startup 1984, 1989) Montgomery, PA "17 "20
Turkey Point, Florida City FL Broward, Dade, 129 (350) 93 (24) q38.6% P-0.08
(2 reactors, startup 1972, 1973) Palm Beach, FL "7 "20
St. Lucie, Hutchinson Island FL Indian River, Martin, 143 (97) 93 (24) q53.8% P-0.04
(2 reactors, startup 1976, 1983) St. Lucie, FL "15 "20
Oyster Creek, Forked River NJ Monmouth, Ocean, NJ 128 (169) 119 (75) q8.1%
(1 reactor, startup 1969) "10 "14
Diablo Canyon, Avila Beach CA San Luis Obispo, 127 (50)b 97 (88) q30.8%
(2 reactors, startup 1984, 1985) Santa Barbara, CA "19 "11
Counting error listed for each sample of teeth. See Appendix Bfor explanation of standard error calculation, Appendix C for
significance testing. Source: US Nuclear Regulatory Commission (www.nrc.gov), obtained August 12, 1999, for reactor locations
and startup dates.
a Average millibecquerels of Sr-90 per gram of calcium.
b In three counties near Limerick, 94 of 98 teeth were from persons born after startup (average 168). In two counties near Diablo
Canyon, 47 of 50 teeth were from persons born after startup (average 135).
Fig. 1. Average Sr-90 in baby teeth, US, by proximity to nuclear plants (persons born 1980–1997).
the next 5 years. Thus, continued decline of Sr-90
from bomb test fallout should have reached a level
approaching zero by about 1980, and averages
should largely represent current sources of this
radionuclide. Average Sr-90 concentration for all
teeth was 132 mBq Sr-90yg Ca, and state averages
ranged from a high of 154 in Pennsylvania to a
low of 108 in California.
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 43
Table 5
Average Sr-90 concentration (by birth year), US, in deciduous teeth (at birth)
Birth year No. teeth Average Sr-90a Counting error
1954–1957 6 191 "78
1958–1961 8 331 "117
1962–1965 8 351 "124
1966–1969 17 272 "66
1970–1973 38 222 "36
1974–1977 38 211 "34
1978–1981 78 140 "16
1982–1985 172 140 "11
1986–1989 532 109 "5
1990–1993 836 132 "5
1994–1997 346 162 "9
% Change, 1986–1989 to 1994–1997 q48.5% P-0.0001
Note: Most teeth are from states of CA, FL, NJ, NY and PA. See Appendix Bfor explanation of error calculation, Appendix C
for significance testing.
a Average millibecquerels of Sr-90 per gram of calcium.
Fig. 2. Average Sr-90 in baby teeth, US, 1954–1997 (mostly CA, FL, NJ, NY, PA).
3.2. By proximity to nuclear reactors
The question of whether those living closest to
nuclear plants have higher burdens of radioactivity
was addressed. Most teeth from residents close to
nuclear plants—defined as counties situated mostly
or completely within 40 miles—include six nuclear
installations, described in Table 4 and Fig. 1.
Average Sr-90 concentrations are compared with
those from all counties in the remainder of the
state, which are farther from reactors.
For each of the six areas, the local average of
Sr-90 exceeded that for the remainder of the state.
Three of the six differences are significant at P-
0.05, with one other of borderline significance
(P-0.08). Aside from a 8.1% excess near the
Oyster Creek plant in central New Jersey, average
Sr-90 concentrations near the other five reactors
ranged from 30.8 to 53.8% above other counties
in these states. Two parts of California can be
considered relatively unexposed control areas. One
is composed of Sacramento and El Dorado, close
44 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
Table 6
Trend in Sr-90 concentration after 1981 in deciduous teeth, at birth by birth year, by state
Birth year No. teeth Average Sr-90ay
counting error
No. teeth Average Sr-90aycounting
error
California Florida
1982–1985 12 104 "31 63 133 "17
1986–1989 50 93 "14 102 112 "11
1990–1993 53 112 "16 192 127 " 9
1994–1997 20 139 "32 99 153 "16
% Change 1986–1989 to 1994–1997 q50.2% q36.3% P-0.04
New Jersey New York
1982–1985 19 117 "27 41 153 "24
1986–1989 71 105 "14 142 120 "10
1990–1993 109 132 "13 237 128 " 9
1994–1997 39 144 "23 104 184 "18
% Change 1986–1989 to 1994–1997 q36.5% q53.6% P-0.002
Pennsylvania All other
1982–1985 6 293 "120 31 134 "24
1986–1989 32 125 "23 135 100 "9
1990–1993 52 152 "21 193 141 "10
1994–1997 36 160 "27 48 159 "23
% Change 1986–1989 to 1994–1997 q27.7% q59.0% P-0.02
See Appendix Bfor explanation of error calculation, Appendix C for significance testing.
a Average millibecquerels of Sr-90 per gram of calcium.
Fig. 3. Average Sr-90 in baby teeth, by state (persons born 1982–1997).
to the Rancho Seco nuclear plant, which closed in
June 1989. The other is the San Francisco Bay
area, which lies approximately 80 miles from
Rancho Seco and 210 miles from the Diablo
Canyon plant. The 50 teeth from persons born
after 1979 near Diablo Canyon have the highest
Sr-90 concentration in the state (127 mBqyg Ca),
followed by those near the closed Rancho Seco
plant (106 mBqyg Ca, 27 teeth), and the San
Francisco Bay area (87 mBqyg Ca, 23 teeth).
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 45
3.3. Temporal trends—total
Temporal trends in average in vivo Sr-90 concentrations
were also analyzed. The earlier St.
Louis study documented a 50% decline in average
Sr-90 concentration in fetal mandibles in the 5
years after the Limited Test Ban Treaty went into
effect (Rosenthal, 1969). The adult bone (vertebrae)
program administered by the US government
showed a similar decline, followed by a more
modest reduction since the mid-1970s; this program
was small in scope, and ceased in 1982
(Klusek, 1984). The teeth analyzed in this report
represent persons born primarily in the 1980s and
1990s, providing data for a population not heretofore
addressed.
Table 5 and Fig. 2 display the trend in average
Sr-90 concentrations from the mid-1950s to the
late 1990s. The trends established by earlier analyses
(a rise until the mid-1960s followed by a
decline until the early 1980s) were duplicated,
even with a limited number of teeth studied prior
to 1980. The new findings for those children born
after 1981, who contributed 91% of all samples in
the study, showed that the decline continued until
the period 1986–1989. Four-year birth cohorts are
used here to maximize numbers of teeth and
smooth trends. In 1986–1989, the lowest average
Sr-90 concentration in the study was observed
(109 mBq Sr-90yg Ca), well below the 351 mBq
Sr-90yg Ca observed in the mid-1960s. This longterm
decline was followed by an increase of 48.5%
in the next two 4-year periods, ending with an
average of 162 mBq Sr-90yg Ca for the 1994–
1997 birth cohort (P-0.0001). Although trends
for individual years are less reliable due to fewer
teeth, the lowest average was reached in 1986 (94
mBq Sr-90yg Ca for 76 teeth) and the highest
average thereafter occurred in 1996 (195 mBq Sr-
90yg Ca for 30 teeth), an increase of 107% (P-
0.007). Only six teeth for births after 1996 have
been analyzed to date.
3.4. Temporal trends—by state
The unexpected and abrupt reversal of declines
in Sr-90 concentration in US baby teeth takes on
greater meaning when data from each state are
analyzed. ‘National’ data essentially include only
five states, and thus may or may not be representative
of the entire US. However, for post-1981
births, each of the five states duplicates the same
trend; a reduction to a post-Test Ban low in 1986–
1989, followed by two successive increases in the
following 4-year periods. The geographic disparity
of these areas suggests that the trend may apply
nation-wide, at least in areas near nuclear reactors,
from which most study teeth were donated. Table
6 and Fig. 3 display these consistent trends, which
also occurred for the ‘all other’ categories (teeth
from children in areas other than the five focus
states). Rises during the 1990s vary from 27.7 to
59.0%. Increases in Florida, New York and ‘other’
states are significant (P-0.05).
3.5. Temporal trends—by counties
The trends in states were also consistent for the
counties (or clusters of counties) that donated the
most teeth to the study (Table 7). These include
MonmouthyOcean County, NJ (closest to the Oyster
Creek plant), Dade County, FL (site of the
Turkey Point plant) and PutnamyRocklandy
Westchester Counties, NY (which converge at the
Indian Point plant). Increases from 1986–1989 to
1994–1997 ranged from 49.8 to 55.7%, with the
Florida and New York counties achieving statistical
significance (P-0.05). The only slight exception
to this trend was that all of MonmouthyOcean
County’s increase took place in the early 1990s.
4. Discussion
The US has conducted no official program
measuring in vivo levels of fission products for
over 20 years. This report introduces current data
on patterns and trends of Sr-90 concentration in
US baby teeth, mostly near nuclear power installations.
The average concentration of Sr-90 was
132 mBq Sr-90yg Ca for all children born after
1979, when in vivo Sr-90 remaining from atomic
46 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
Table 7
Trend in Sr-90 concentration after 1981 in deciduous teeth (at birth) by birth year, by county (counties with the largest sample
sizes)
Birth year No. teeth Average Sr-90a Counting error
Dade County FL
1982–1985 47 141 "21
1986–1989 57 94 "13
1990–1993 106 124 "12
1994–1997 43 141 "22
% Change 1986–1989 to 1994–1997 q50.6% P-0.057
Monmouth, Ocean Counties NJ
1982–1985 13 150 "40
1986–1989 44 93 "14
1990–1993 76 140 "16
1994–1997 31 139 "25
% Change 1986–1989 to 1994–1997 q49.8%
Putnam, Rockland, Westchester Counties NY
1982–1985 17 202 "50
1986–1989 43 135 "21
1990–1993 101 148 "15
1994–1997 52 211 "30
% Change 1986–1989 to 1994–1997 q55.7% P-0.04
See Appendix Bfor explanation of error calculation, Appendix C for significance testing.
a Average millibecquerels of Sr-90 per gram of calcium.
bomb tests should approach 0.5 This concentration
is lower than that in those born before 1980, when
bomb test fallout accounted for a substantial proportion
of in vivo radioactivity. However, it
exceeds the levels before the large-scale testing
began in 1951 in Nevada (Rosenthal, 1969).
Long-term declines first slowed in the 1982–
1985 period, when no change was observed from
the previous 4 years. The reason(s) for this departure
is not certain. The decline resumed into the
period 1986–1989.
The most dramatic and unexpected finding in
this report is the reversal after the late 1980s of
decades-long declines in average Sr-90 concentra-
5 Stamoulis et al. (1999) contains a chart summarizing
trends in Sr-90 in deciduous teeth from various European
nations and the Soviet Union. The chart shows that, from a
level of approximately 10 mBqyg Ca in 1951, a peak of 250
was reached in 1964, similar to the US trend. By 1975, the
average level had fallen to approximately 30 (three times the
1951 average) and was still declining. Three times the 1951
US average of just over 7 means that the 1975 US Sr-90
average should have been approximately 22. But the actual
1975 average found by RPHP was 183 (12 teeth), and 198
for 29 teeth from 1974–1976 births.
tion. We observed a 48.5% higher concentration
in 1994–1997 births over 1986–1989 births (162
vs. 109 mBq Sr-90yg Ca), a trend consistent for
each of five states (and counties in these states
near nuclear reactors) included in this study. This
temporal change cannot represent the continued
decay of old bomb test fallout from Nevada; rather,
it probably represents rising amounts of a currently
produced source of environmental radioactivity
entering the body. Current sources of Sr-90, a manmade
fission product, are limited during the 1990s,
and most are not likely to account for recently
rising levels of Sr-90 in baby teeth.
(1) Fallout from the 1986 Chernobyl accident
(including Sr-90) entered the US environment,
raising levels of long-lived radionuclides, but these
returned to pre-1986 levels within 3 years (Mangano,
1997; National Air and Radiation Environmental
Laboratory, 1975–2001). For example, a
rise of 98–311 mBq Cesium-137yl in pasteurized
milk occurred in 60 US cities from May–June
1985 to May–June 1986, when Chernobyl fallout
levels in the US peaked. This concentration in the
same 2-month period in the following years
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 47
declined to 242, 155 and 81 mBq Cs-137yl,
returning to pre-Chernobyl levels in 1989. Because
Cs-137 has a half-life (30 years) similar to Sr-90
(28.7 years), it is logical that environmental (and
thus, in vivo) Sr-90 from Chernobyl followed the
same general pattern.
Another factor suggesting Chernobyl fallout
probably does not account for the fact that post-
1989 Sr-90 increases in baby teeth is the consistent
finding of higher Sr-90 concentrations near nuclear
power plants. Chernobyl fallout levels varied by
geographic area, with the northwest US (where
there is only one nuclear power reactor, in Washington
state) receiving the highest level of radionuclide
deposits.
(2) The increase probably does not represent
high-level nuclear waste generated by reactors,
which is generally stored in deep pools of cooled
water or in casks below or above ground. Despite
the leakage of some casks, the radioactivity contained
in the waste is currently not in the food
chain.
(3) Academic-based research reactors also produce
fission products. However, these reactors are
small in size and few (and declining) in number,
which makes it an unlikely reason accounting for
such a widespread and sustained trend in Sr-90 in
bodies.
(4) Nuclear submarines produce fission products,
but they are either contained within the
submarine or released into the ocean. Thus, this is
not a source of Sr-90 in the food chain, and not a
reason for the rise documented in this report.
(5) Emissions from nuclear weapons plants
account for another source of Sr-90. However, all
reactors involved with producing nuclear weapons
ceased manufacturing operations by 1991, and are
not likely to play a role in rising Sr-90 concentrations
after that time.
(6) While the last above-ground atomic bomb
test took place in 1962, subterranean tests at the
Nevada Test Site continued. Some of these tests
vented radioactivity into the atmosphere. These
emissions were much smaller than the atmospheric
tests, and the last such test occurred in September
1992 (Norris and Cochran, 1994), making it an
unlikely contributor to increases in Sr-90 throughout
the 1990s.
(7) Reprocessing of nuclear fuels also creates
fission products, but was ceased in the US in the
late 1970s, and is not a factor in recent rises in
Sr-90.
The only other source of Sr-90 that can explain
this steady and dramatic rise in the 1990s is
emissions from nuclear power reactors. Because
reactors operated a greater percentage of the time,
average annual generation of electricity rose 37.5%
from 475 000 to 653 000 GW h from 1986–1989
vs. 1994–1997, an increase not markedly different
from the 48.5% rise in average Sr-90 levels at
birth (US Nuclear Regulatory Commission, 2001).
Determining the extent of the correlation between
these two trends requires more precise
investigation.
Another major finding is that the counties located
within 40 miles of each of six nuclear reactors
have consistently higher Sr-90 levels than other
counties in the same state. These counties were
selected to generally correspond with those used
by the US National Cancer Institute in a study of
cancer near nuclear plants (Jablon et al., 1990).
The excess near each nuclear plant ranged, with
one exception, from 30.8 to 53.8% higher. More
study, assessing whether locally produced radioactivity
entering the body from inhalation andyor
locally produced food and water account for these
consistent differences, is merited. Findings on
doses near reactors should be compared with health
data. For example, childhood cancer rates near 14
of 14 eastern US reactors exceed the national rate
(Mangano et al., 2003).
This analysis of proximity arrives at a different
conclusion than an earlier report (O’Donnell et al.,
1997) that found no correlation between distance
from the Sellafield nuclear plant in western England
and Sr-90 levels in baby teeth. That study
used a regression equation to test this relationship.
There are methodological and analytical differences
between the two studies. O’Donnell considered
teeth from as far as 300 miles from Sellafield,
without taking into account Sr-90 produced by
reactors other than Sellafield, while this report
used only the counties most proximate (within 40
miles) to reactors. That report tested teeth in
batches, while this study used individual readings.
Factors other than distance from the radiation
48 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
source may influence Sr-90 levels in vivo. The
uptake of radioactivity in fetal tooth buds depends
on intake during pregnancyyearly infancy and
transfer from maternal bone stores, which vary
from person to person. These in turn can be
dependent on food and water source, along with
dietary differences.
A third major finding is that average Sr-90
concentrations vary geographically. Children from
Pennsylvania (mostly near Pottstown, close to
Philadelphia) who donated teeth had the highest
average Sr-90 of the five states studied. Pottstown
lies within 70 miles of 11 operating (and 2 closed)
reactors, a concentration unmatched in the US.
California, especially areas not close to nuclear
reactors, is the state with the lowest average Sr-
90. There are only four nuclear reactors on the
entire west coast in operation since 1992, compared
to dozens in the northeast.
At present (pending more detailed study), nuclear
power reactors appear to be the most likely
source explaining the recent unexpected rise in Sr-
90 concentrations, and elevated Sr-90 levels nearest
the plants. The geographic consistency and
longevity of these trends and patterns, plus the
large number of teeth studied, make these patterns
meaningful and (in many instances) statistically
significant. The fact that gross beta in US precipitation
continued to rise after 1997 and that the
highest average Sr-90 level since a low point was
reached in 1986 occurred in the most recent birth
year studied (1996, 195 mBq Sr-90yg Ca in 30
teeth) suggest that this trend may continue in the
near future.
5. Study limitationsyopportunities for further
study
This report represents the first large-scale study
of US in vivo levels of radioactivity in several
decades. Although the initial findings presented
here are important ones, they raise various questions
that should be addressed in future research.
Other unexplored factors may help explain the
temporal trends affected here. For example, the
current study collected auxiliary data on mother’s
age at delivery and source of drinking water.
Analyzing results by basic characteristics such as
gender and race can be performed in future studies.
Some factors that affect in vivo levels are already
known. For example, children who are breast-fed
accumulate lower Sr-90 concentrations than do
bottle-fed infants (Rosenthal, 1969). Other dietary
differences and their effects on Sr-90 levels can
be further explored in future research.
Despite the consistency of results across geographic
areas, substantial numbers of teeth were
tested from only 5 of 50 US states. More teeth
from other states would enhance knowledge about
recent patterns of in vivo radioactivity. For example,
19 of the 50 US states (many in the western
US) have no operating nuclear reactors, and may
display patterns of Sr-90 different than the five
already analyzed. The comparison could be extended
to nations with no operating nuclear reactors
(such as the Philippino teeth mentioned in this
report). Testing the hypothesis that these states
have lower levels of Sr-90 would be appropriate
and necessary in future reporting of results.
The study did not collect sufficient teeth to
compare local Sr-90 levels before and after a
nuclear reactor opens. The hypothesis that opening
a reactor will raise average in vivo concentrations
and closing a reactor will reduce them should be
tested.
A potential follow-up to this report is to institute
a public program measuring in vivo levels in
humans andyor animals near nuclear plants for the
first time. In addition, more radionuclides in the
environment (air, water, soil, etc.) may be tracked.
The US government maintains such records near
nuclear plants, but has phased out public reporting
of several isotopes and failed to perform any longterm
analysis.
The data presented herein describe past and
current patterns of radioactivity in children’s teeth.
The three in vivo programs of measuring Sr-90 in
US teeth and bones were never accompanied by
any reports assessing potential health risks from
this radioactivity. The current tooth study previously
documented that average Sr-90 levels and
childhood cancer rates followed similar trends
during atmospheric bomb testing in the 1950s and
1960s. In addition, on Long Island, New York,
recent Sr-90 trends correlate closely to trends in
childhood cancer incidence, after a 3-year latency
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 49
period (Gould et al., 2000a). Thus, comparing
radioactivity and health patterns should be central
to any follow-up of this analysis.
Acknowledgments
Jerry Brown, Ph.D., is acknowledged for his
contribution in collecting baby teeth in southeastern
Florida.
Appendix A: Determination of Sr-90 to calcium
ratio
Sr-90 in deciduous teeth was determined under
the direction of Hari D. Sharma, Professor Emeritus
of Radiochemistry and president of REMS,
Inc., Waterloo, Ontario, Canada. Employing a
1220-003 Quantulus Ultra Low-Level Liquid Scintillation
Spectrometer manufactured by the Perkin-
Elmer Company in Massachusetts, Dr Sharma
followed the following procedure.
Water-washed teeth were treated with 30%
hydrogen peroxide for a period of 24 h to ensure
that organic material adhering to teeth was oxidized.
Teeth were then scrubbed with a hard brush
for removing oxidized organic material and the
fillings. Teeth were then dried at 110 8C for several
minutes and then ground to a fine powder (ball
mill). It is very important to remove any filling
because if left behind inside a tooth, it tends to
give colored solution or dissolution in a mineral
acid. The presence of colored solution reduces the
efficiency of counting.
Approximately 0.1 g of the powder is weighed
in a vial, then digested for a few hours with 0.5
ml of concentrated nitric acid along with solutions
containing 5 mg of Sr2q and 2 mg of Y3q carriers
at approximately 110 8C on a sand bath. The
solution is not evaporated to dryness. The digested
powder is transferred to a centrifuge tube by
rinsing with tritium-free water. Carbonates of Sr,
Y and Ca are precipitated by addition of a saturated
solution of sodium carbonate, and then centrifuged.
The carbonates are repeatedly washed with a dilute
solution of sodium carbonate to remove any coloration
from the precipitate. The precipitate is
dissolved in hydrochloric acid, and the pH is
adjusted to 1.5–2 to make a volume of 2 ml, of
which 0.1 ml is set aside for the determination of
calcium. The remaining 1.9 ml is mixed with 9.1
ml of scintillation cocktail Ultima Gold AB, supplied
by Packard Bioscience BV in a special vial
for counting. A blank with appropriate amounts of
Ca2q, Sr2q and Y3q is prepared for recording the
background.
The activity in the vial with the dissolved tooth
is counted four times, 100 min each time, for a
total of 400 min, with the scintillation spectrometer,
to improve accuracy of results. The background
count-rate in the 400–1000 channels is 2.25"0.02
countsymin. The background has been counted for
over 5000 min so that the error associated with
the background measurement is approximately 1%.
The overall uncertainty or one sigma associated
with the measurement of Sr-90 per gram of calcium
is "26 mBqyg Ca.
The efficiency of counting was established using
a calibrated solution of Sr-90yY-90 obtained from
the National Institute of Standards and Technology,
using the following procedure. The calibrated solution
is diluted in water containing a few milligrams
of Sr2q solution, and the count-rate from an aliquot
of the solution is recorded in channel numbers
ranging from 400 to 1000 in order to determine
the counting efficiency for the beta particles emitted
by Sr-90 and Y-90. It is ensured that the Y-90
is in secular equilibrium with its parent Sr-90 in
the solution. The counting efficiency was found to
be 1.67 counts per decay of Sr-90 with 1.9 ml of
Sr-90yY-90 solution with 25 mg of Ca2q, 5 mg of
Sr2q, 2 mg of Y3q and 9.1 ml of the scintillation
cocktail.
The calcium content was determined by using
an Inductively Coupled Plasma instrument. The
analysis is provided to REMS, Inc., by the University
of Waterloo laboratories. REMS is located
at P.O. Box 33030, Waterloo, Ontario, Canada,
N2T2M9.
Appendix B: Calculation of counting error for
Sr-90 in baby teeth due to laboratory observation
and sample size
In Tables 3–7, the counting error for average
concentrations of Sr-90 is calculated for each state
as a combination of two variables: the error due
50 J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51
to laboratory observation and the error due to
sample size. Calculating each of these errors are
as follows, using all 133 teeth (average mBq Sr-
90yg Cas155) from Pennsylvania as an example.
These data appear in Table 3.
Lab observation: The count of mBq of Sr-90 is
not an exact one, but carries an uncertainty due to
limitations of the counter. The error range for an
individual tooth is "26 mBq, a conservative
estimate that may be lower for teeth with larger
mass. Thus, the lab observation error for a sample
of 133 teeth is
26 mBqy6Ns26 mBqy6133s2.25 mBq
Sample size: The error due to the sample size is
1y6Ns1y6133s13.44 mBq
Calculation: The squares of the two results are
added quadratically. Thus,
6((2.25)2q(13.44)2)
s13.63 mBq (rounded to 14)
With an average Sr-90 concentration for the 133
teeth of 155 mBqyg Ca, the confidence interval is
between 127 and 183, or 155 plus or minus 28 (2
times 14). Thus, there is a 95% chance that the
actual average of the entire population falls within
127 and 183.
Appendix C: Calculation of significance of differences
in Sr-90 averages between counties
near reactors and more distant counties
In Table 4, average Sr-90 concentrations in teeth
from counties near nuclear reactors were compared
with averages from other counties in the same
state. The significance of differences between the
two means was calculated using a t-test.
For example, the mean Sr-90 concentration for
counties closest to the Indian Point reactor was
164 mBqyg Ca (217 teeth), compared to 121 (317
teeth) for other counties in New York State. The
formula used for the significance of this difference
is as follows:
Counties near Indian point: "{1y6217}=164
s11.1 (rounded to 11)
Other counties in New York state: "{1y6317}
=121s6.8 (rounded to 7)
{164y121}y6(112q72)s(45y13.04)s3.45
In a basic statistics table, 3.45 standard deviations
(z score) indicate a P value of -0.001, i.e.
there is less than a 1 in 1000 chance that the
difference is due to random chance.
In Tables 5–7, the significance of differences in
average Sr-90 concentrations from 1986–1989 to
1994–1997 were tested using a similar technique.
For example, using Florida data in Table 6
1986y1989; for 102 teeth,
average mBq Sr-90yg Cas112
1994y1997; for 99 teeth,
average mBq Sr-90yg Cas153
1986y1989s"?1yy102?=112s11.1
(rounded to 11)
1994y1997s"?1yy99?=153s15.4
(rounded to 15)
{153y112}y6(112q152)s2.20
In a basic statistics table, 2.20 standard deviations
(z score) indicate a P value of -0.04, i.e.
there is less than a 4 in 100 chance that the
difference is due to random chance.
References
Aarkrog A. Strontium-90 in shed deciduous teeth collected in
Denmark, the Faroes, and Greenland from children born in
1950–1958. Health Phys 1968;15:105.
Baratta EJ, Ferri ES, Wall MA. Strontium-90 in human bones
in the United States, 1962–1966. Radiol Health Data
1970;April:183 –186.
Fracassini C. The cancer time bomb facing Scots born during
Cold War. A January 20, 2002 article, obtained January 21,
2002 at http:yynews.scotsman.comyscotland.cfm.
Gould JM, Sternglass EJ, Sherman JS, Brown JB, McDonnell
W, Mangano JJ. Strontium-90 in deciduous teeth as a factor
J.J. Mangano et al. / The Science of the Total Environment 317 (2003) 37–51 51
in early childhood cancer. Int J Health Serv 2000;30:515 –
539.
Gould JM, Sternglass EJ, Mangano JJ, McDonnell W, Sherman
JD, Brown J. The strontium-90 baby teeth study and
childhood cancer. Eur J Oncol 2000;5:119 –125.
Jablon S, Hrubec Z, Boice JD, Stone BJ. Cancer in Populations
Living Near Nuclear Facilities. National Cancer Institute:
NIH Publication 90-874. Washington DC: US Government
Printing Office; 1990.
Kalckar HM. An international milk teeth radiation census.
Nature 1958;August:283 –284.
Klusek CS. Strontium-90 in human bone in the US, 1982.
New York: US Department of Energy, 1984, Report No.
EML-435, p. 1–27.
Kohlehmainen L, Rytomaa I. Strontium-90 in deciduous teeth
in Finland: a follow-up study. Acta Odontol Scand
1975;33:107 –110.
Kulev YD, Polikarpov GG, Prigodey EV, Assimakopoulos PA.
Strontium-90 concentrations in human teeth in South
Ukraine, 5 years after the Chernobyl accident. Sci Total
Environ 1994;155:214 –219.
Mangano JJ. Childhood leukaemia in US may have risen due
to fallout from Chernobyl. BMJ 1997;314:1200.
Mangano JJ, Sternglass EJ, Gould JM, Sherman JD, Brown J,
McDonnell W. Strontium-90 in newborns and childhood
disease. Arch Environ Health 2000;55:240 –245.
Mangano JJ, Sherman J, Chang C, Dave A, Feinberg E, Frimer
M. Elevated childhood cancer incidence proximate to U.S.
nuclear power plants. Arch Environ Health 2003;58(2):1 –
8.
National Air and Radiation Environmental Laboratory. Environmental
Radiation Data Report. Montgomery AL: US
Environmental Protection Agency, quarterly volumes, 1975–
2001.
Norris RS, Cochran TB. United States nuclear tests, July 1945
to 31 December 1992. Washington: Natural Resources
Defense Council, 1994. p. 58.
O’Donnell RG, Mitchell PI, Priest ND, Strange L, Fox A,
Henshaw DL, Long SC. Variations in the concentration of
plutonium, strontium-90 and total alpha-emitters in human
teeth collected within the British Isles. Sci Total Environ
1997;201:235 –243.
Rosenthal HL. Accumulation of environmental strontium-90
in teeth of children. In Proceedings of the Ninth Annual
Hanford Biology Symposium at Richland Washington, May
5–8, 1969. Washington DC: US Atomic Energy Commission;
1969.
Rosenthal HL, Austin S, O’Neill S, Takeuchi K, Bird JT,
Gilster JE. Incorporation of fall-out strontium-90 in deciduous
incisors and foetal bone. Nature 1964;August:616 –
617.
Rytomaa I. Strontium-90 in deciduous teeth collected in
northern Finland from children born in 1952–1964. Acta
Odontol Scand 1972;30:219 –233.
Santholzer W, Knaifl J. Strontium-90 content of deciduous
human teeth. Nature 1966;212:820.
Scholz R. Ten years after Chernobyl: The rise of strontium-90
in baby teeth. Munich: The Otto Hug Radiation Institute,
distributed with permission from the German branch of the
International Physicians for the Prevention of Nuclear War;
1996.
Stamoulis KC, Assimakopoulos PA, Ioannides KG, Johnson
E, Soucacos PN. Strontium-90 concentration measurements
in human bones and teeth in Greece. Sci Total Environ
1999;229:165 –182.
US Nuclear Regulatory Commission: Information Digest 2000
Edition, Available from: www.nrc.gov, May 7, 2001.