None of the
patients who had died had
bone- or liver-related malignancies as
the listed (or even the contributing)
cause of death on their death certifi-
cates, unless that was the diagnosed
disease at the time of the injection.
And those patients who were still living
also did not show any plutonium-
related effects.
Eight of the 18 cases had survived at
least twice as long as the four-year pe-
riod established as the shortest induc-
tion interval for a radium-induced bone
tumor. Using known cases of bone tu-
mors from radium, Rowland and
Durbin estimated that “the lowest aver-
age endosteal [bone surface] dose at
which plutonium might induce bone tu-
mors in man to be of the order of 600
rad.” Four of the patients injected with
plutonium had considerably higher en-
dosteal doses (7420, 1280, 1790, and
973 rad); the other four had significant
fractions of that dose (141 to 448 rad).
Although, one to three cases of bone
cancer were possible in the group, none
had appeared (which might indicate a
higher threshold dose for bone cancer
or simply be a result of the smallness
of the group). In regard to doses to the
liver, all but one of the cases had esti-
mated doses that were smaller than
what appeared necessary, in comparison
to radium, to cause liver cancer. Thus,
it was not surprising that no liver tu-
mors had appeared.
A Recent Analysis of the
Excretion Data
One outcome of the openness inititative
pledged by the Department of Energy
and the subsequent review of docu-
ments was a re-analysis of the plutoni-
um injection data by one of the authors
(Moss) and Gary Tietjen. A careful re-
view of the original notebooks at Los
Alamos has revealed some errors in the
urinary excretion data for the Rochester
patients. Some of those errors were
mistakes, others were simply needed
adjustments for chemical recovery and
elapsed collection time. For example,
failure in the Rochester metabolic ward
to properly time the urine sampling
from the time of injection led to uncer-
tainties in the initial excretion rates.
Likewise, some of the data were not
corrected for the analytically measured
per cent recovery of plutonium, includ-
ing an 88-per-cent recovery rate of plu-
tonium for all the Rochester urine data.
When there was insufficient information
to check the values, Moss and Tietjen
discarded the data. In many cases,
however, careful documention allowed
the original data to be corrected and in-
cluded in the subsequent analysis.
(After 1956, a different urinalysis pro-
cedure, based on a nuclear-track
method developed at Hanford, was im-
plemented at Los Alamos, and data
from that time onward are much more
accurate and consistent. Today’s ana-
lytical methods routinely detect body
burdens at the 0.1-microgram level.)
As a result of the re-examination of
original data, it is apparent that the in-
crease in excretion rate noted by Rundo
was, in fact, only an artifact, the result
of urine assays that were not corrected
for chemical yield or for alpha-counting
instrument calibration bias.
Also included in the re-analysis were
several consecutive daily samples that
had been collected from each of HP-3,
HP-4, HP-6, and HP-9 about a year
after their injections. Although these
data were recorded at Los Alamos, for
some unknown reason Langham may
not have been aware of them; they were
not used in his analysis even though
they were consistent with the data he
did use (the 500-day data obtained from
HP-6).
In addition to corrections, new data
have become available from a recent
study. Talbot, Newton, and Warner in
England injected plutonium-237 into
two healthy male volunteers and ana-
lyzed the excreta using modern analyti-
cal methods. Plutonium-237 has only a
45.3-day half-life and decays by the rel-
atively benign electron-capture mode,
which made this isotope a negligible
health concern compared to plutonium-
239. Moreover, x rays emitted in the
decay enabled patterns of organ uptake
to be studied during the experiment.
This approach was not used earlier be-
cause it has been too difficult to elimi-
nate other plutonium isotopes with long
half-lives. In this case, the researchers
were able to use a variable-energy cy-
clotron at Harwell and adjust the condi-
tions of the irradiation of uranium-235
with helium ions to make relatively
pure plutonium-237.
Moss and Tietjen used the new excre-
tion data together with the corrections
to the original plutonium-239 data to do
another analysis of plutonium urinary
excretion. Based solely on empirical
grounds, they expanded Langham’s
original power function by adding a
second term. The urine data for the
two plutonium-237 subjects from day 5
through day 15 are remarkably linear
on a log-log plot, whereas the data for
days 1 through 4 are more variable.
Thus, only the data for days 5 through
14 were used to obtain the first power-
function term. When they compared
the slope for that term to the slopes for
ten of the Rochester patients (HP-1
through HP-10), the comparison, for the
most part, was very close.
Moss and Tietjen next used the sparse
The Human Plutonium Injection Experiments
Number 23 1995 Los Alamos Science
219
Bill Moss and Eileen Welsome