The British Journal of Radiology, 76 (2003), 246–253 E 2003 The British Institute of Radiology

DOI: 10.1259/bjr/82373369

Initial medical management of patients severely irradiated in

the Tokai-mura criticality accident
1
T HIRAMA, MD, PhD, 1S TANOSAKI, MD, PhD, 2S KANDATSU, MD, PhD, 1N KUROIWA, MD,
2
T KAMADA, MD, PhD, 2H TSUJI, MD, PhD, 2S YAMADA, MD, PhD, 2H KATOH, MD, PhD,
2
N YAMAMOTO, MD, PhD, 2H TSUJII, MD, PhD, 3G SUZUKI, MD, PhD and 1M AKASHI, MD, PhD
1
Research Center for Radiation Emergency Medicine, 2Research Center for Charged Particle Therapy, National Institute
of Radiological Sciences, Chiba, Japan and, 3Radiation Effect Research Foundation, Hiroshima, Japan

Abstract. A nuclear criticality accident occurred in Japan on September 30, 1999, which resulted in severe
exposure of three victims to mixed flux of neutrons and c-rays. Estimated average doses for the three victims
were 5.4 Gy of neutrons and 8.5 Gy of c-rays for Patient A, 2.9 Gy of neutrons and 4.5 Gy of c-rays for Patient
B, and 0.81 Gy of neutrons and 1.3 Gy of c-rays for Patient C. They then suffered the consequences of the
effects of ionizing radiation resulting in acute radiation syndrome. In Patients A and B, bone marrow failure
was so severe that they received haematopoietic stem cell transplantation. The graft initially took successfully in
both patients, although in Patient B it was later taken over by his own haematopoietic cells. They also suffered
from severe skin lesions, later exhibited gastrointestinal bleeding and eventually died of multiple organ failure
82 and 210 days after the accident, respectively. The survival of these patients beyond the period of
agranulocytosis means that bone marrow failure per se caused by exposure to ionizing radiation may now be
overcome. Patient C also developed bone marrow failure and was treated with granulocyte colony-stimulating
factor as well as supportive care. He recovered without major complications and is now under periodical follow-
up. Remarkably, during the prodromal phase, all the patients exhibited hypoxaemia, two of whom also showed
interstitial oedema of the lungs. In Patient C these manifestations improved within a week. The circumstances of
the accident and the initial medical treatment of the victims are described.

Since the advent of the nuclear industry, more than 40
accidents associated with nuclear criticality in undesired
manners have occurred [1, 2]. Most occurred before 1970,
on which our knowledge of the medical outcomes of such
accidents is mainly based. Criticality accidents differ from
other types of radiation accidents in that they cause
exposure to neutrons and c-rays in quite a short period,
often a moment, and can be lethal for those within several
metres of the source of radiation, who are often those who
triggered the chain reaction. Those who were in the
immediate vicinity of the source of radiation exhibit symp-
toms reflecting the severe damaging effects of high-dose
ionizing radiation to many organ systems, such as the bone
marrow, gastrointestinal tract, cardiovascular system and
skin, which are collectively described as a severe form of
acute radiation syndrome (ARS). Typically, survival terms
of such patients after a criticality accident have been less
than 10 days [3, 4]. For instance, in a reported criticality
accident in 1946 in the United States [5], the victim who
was directly handling the critical assembly survived 9 days.
In another recent criticality accident in Russia, the victim
survived 66.5 h [6]. Earliest clinical manifestations in such
patients, following non-specific prodromal symptoms such
as nausea and vomiting, have been early and severe leuko-
penia and painful oedema of the upper extremities, which
in most cases had been closest to the source of radiation.
Received 12 November 2001 and in revised form 4 October 2002,
accepted 11 November 2002.
Address correspondence to Toshiyasu Hirama, Research Center for
Radiation Emergency Medicine, National Institute of Radiological
Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan.
Thus, for the survival chance of patients with a severe
form of ARS to ever exist, therapeutic intervention to
support general physical condition as well as bone marrow
failure is of the utmost importance. In contrast, in the
survivors of previous criticality accidents, radiation-
induced skin damage has typically been absent, except
localized epilation, and the decrease in leukocytes has been
slower and milder.
In the criticality accident in Tokai-mura, Japan, which
occurred in 1999, three victims were exposed to high doses
of neutrons and c-rays. Two of the victims exhibited
extremely severe damage to the bone marrow as well as to
the skin, which was comparable with a severe form of
ARS. However, state of the art therapy, including
haematopoietic stem cell transplantation, enabled them
to survive 82 days and 210 days after the accident,
illustrating that bone marrow failure per se may no
longer be the limiting factor for the survival of such
patients. In the present report, we describe the circum-
stances in which the victims were irradiated, their medical
condition and the therapeutic strategy in the initial period.
We also show that arterial blood gas analysis and CT of
the chest may be of diagnostic value, at least in criticality
accidents, even when the symptoms of ARS are still subtle.
Patients and methods
Patients
Patients A, B and C, all male, were 35 years old,
39 years old and 54 years old, respectively, at the time of
the accident. Disclosure of any information related to

246 The British Journal of Radiology, April 2003

Tokai-mura criticality accident
the patients in this report is based on written informed
consent.
CT
CT was performed with a HighSpeed Advantage (GE
Medical Systems, Milwaukee, WI) scanner. Scans were
obtained at end inspiration using 10 mm collimation and
at 10 mm intervals through the chest, with the patient
in a supine position. CT was performed at 140 kVp and
200 mAs.
The accident
A nuclear criticality accident occurred in Tokai-mura,
located approximately 120 km northeast of Tokyo. It
occurred in a company that had been manufacturing
nuclear fuel for power plants in Japan. On September 30,
1999, three workers were mixing together seven batches
(2.4 kg uranium per batch) of individually prepared uranyl
nitrate solution, one of the final steps in the process of
producing 18.8% uranium for an experimental fast reactor
in Tokai-mura. Previously they had used a container with
a controlled geometry for this purpose to prevent criti-
cality. On that day, however, they decided to mix the
batches in a larger container, the precipitation tank, of
which the mass limit to prevent criticality was 2.4 kg
uranium. At 10:35 am on September 30, when they were
pouring the seventh batch of solution into the tank and its
uranium content reached approximately 16 kg, criticality
was triggered. The positions and postures of the workers
at the scene were reconstructed by interviewing them
(Figure 1). Worker A, who was the most severely irra-
diated, apparently was standing beside and facing the tank
holding a funnel with at least his right hand. Worker B,
who was also severely irradiated, was crouching on a stage
with his right foot on a stair, pouring the solution from a
bucket into the tank, with his face and right hand very
close to the tank. Worker C, who was less severely
irradiated, was sitting at his desk a few metres away from
the tank separated by a thin wall. At the moment that
Figure 1. The positions and postures of the victims of the
accident at the moment when criticality was triggered, recon-
structed by interviewing Workers B and C.
The British Journal of Radiology, April 2003
criticality was triggered, all three workers saw a flash of
blue light. Workers A and B then ran to the next building,
where Worker A collapsed and lost consciousness for some
20 s. He then experienced vomiting and diarrhoea. The
stools were not described. Worker C stayed at the scene
for approximately 5 min trying to contact the radiation
safety personnel of the facility by telephone. He then
joined the other two workers. The three workers (hereafter
referred to as patients) were initially transported to
National Mito Hospital. Because the patients were radioac-
tive by routine radiological survey and their granulocytosis
and lymphopenia were already marked, physicians of the
hospital readily recognized the severity of the circum-
stances and decided to send the patients to the National
Institute of Radiological Sciences (NIRS) in Chiba, which
had been assigned, as the hospital in charge of medical
treatment of victims of radiation accidents in Japan’s Basic
Plans for Disaster Prevention. The three patients arrived at
the NIRS at 15:10. While the physicians of the NIRS were
evaluating the medical condition of the patients, our
health physicists and radiation safety staff made radi-
ological analyses. Their body surface was significantly
radioactive as surveyed by Geiger–Muller survey meter,
but not by alpha survey meter. The vomitus of Patient A
was analyzed and 24Na detected [7], which is produced by
neutron activation of stable 23Na and emits b- and c-rays
and has been routinely utilized to calculate neutron flux to
estimate doses in previous criticality accidents. It was thus
known that these were victims of a criticality accident.
Evaluation of patients
Within hours of total body irradiation, radiation acci-
dent victims exhibit symptoms such as nausea, vomiting,
diarrhoea, fever and in severe cases, hypotension and even
loss of consciousness, which might reflect an acute
response of the neurovascular system [8]. Although none
of these symptoms are specific for ARS, collectively their
promptness and severity have been described to reflect the
extent of damage that the body had received from ionizing
radiation and have often been compatible with the dose of
radiation estimated by physical methods [9]. Therefore, we
carefully observed the initial symptoms and physical signs
of the patients to get a rough picture of the severity of the
effect of irradiation and thereby formulate reasonable
therapeutic strategies [10].
Patient A experienced nausea, vomiting and a transient
loss of consciousness minutes after the exposure, and
diarrhoea within 1 h. The type of stools at the accident
scene was not reported, but during his 3-day admission at
the NIRS, stools were watery and negative for occult
blood, which meant that the intestinal epithelium was still
intact. On admission, the patient was febrile without any
evidence of infection, slightly drowsy and hypotensive with
a systolic blood pressure of 70 mmHg. The patient also
had diffuse erythema on the ventral body surface, facial
oedema, injection of the conjunctiva bulbi and painful
bilateral parotid swelling. He complained of diffuse
tenderness of the abdominal wall by palpation and dif-
ficulty in voiding. These findings strongly indicated that
the patient had received serious damage from the irra-
diation, comparable with the victims of reported accidents
with fatal outcome.
Patient B also experienced nausea and vomiting within
247
 T Hirama, S Tanosaki, S Kandatsu et al

1 h of exposure, but without early diarrhoea. Although
Patient B was normotensive on the day of the accident, his
blood pressure was rather low for the next several days
(lowest recorded 80/44 mmHg). The patient was also
slightly drowsy, febrile, had erythema on the ventral body
surface and salivary gland swelling, and complained of
mild epigastralgia on admission. These findings indicated
that he would also undergo a severe form of ARS,
although to a lesser degree than Patient A.
Patient C experienced only mild nausea several hours
after exposure. On admission, his skin had slight but
diffuse erythema. Otherwise, the patient was asymptomatic
and appeared to be in a good physical condition. If we
utilize the published table by the International Atomic
Energy Agency that correlates symptoms of ARS during
its prodromal phase with results of dosimetry, the symp-
toms and signs in Patients A, B and C would correspond
to the degrees of damage caused by more than 8 Gy,
between 6 and 8 Gy and less than 4 Gy, respectively, of
ionizing radiation [9].
Lymphocytes represent one of the most sensitive types
of cells to ionizing radiation; the rapidity and extent of the
decrease of lymphocytes are known to correlate with the
severity of exposure. Baranov et al have formulated a
function between the lymphocyte count and the corre-
sponding exposure dose in ARS [11]. Although the method
can be influenced by the baseline numbers of lymphocytes,
which vary from one individual to another, it provides
information about the severity of ARS in the first
24 h, when the results of dosimetry are typically not yet
available. Because their data were based on c-ray accidents,
we plotted the patients’ lymphocyte counts in the formula
and obtained values equivalent to c-rays. The resultant
values for Patients A, B and C were more than 10 Grey
Equivalent (GyEq), between 6 GyEq and 10 GyEq, and
between 1 GyEq and 4.5 GyEq, respectively. In Patient A,
the lymphocyte counts were far outside the range of
the graph.
The results of dosimetric analyses by specialists came
2 days after the accident and were refined thereafter. They
agreed fairly well with the above values. Doses based on
the 24Na content in the peripheral blood were 5.4 Gy of
neutrons and 8.5 Gy of c-rays for Patient A, 2.9 Gy of
neutrons and 4.5 Gy of c-rays for Patient B and 0.81 Gy
of neutrons and 1.3 Gy of c-rays for Patient C 12].
Chromosome analyses of the patients’ lymphocytes, utili-
zing the prematurely condensed ring method, resulted in
doses equivalent to c-rays, which were more than 20 GyEq
for Patient A, 7.4 GyEq for Patient B and 2.3 GyEq for
Patient C [13].
Laboratory analyses revealed several remarkable find-
ings (Table 1a). All three patients showed granulocytosis
and degrees of lymphocytopenia on the day of the accident
(day 0). Serum amylase increased after admission and
peaked on the day following the accident (day 1). It then
decreased and, in Patient C, returned to within the normal
range on day 4. Isoenzyme analyses of serum amylase
revealed a predominant S-fraction, indicating damage to
the salivary glands. Serum uric acid also increased in
Patients A and B on day 1. Arterial blood gas was
analyzed in all of them and invariably showed hypox-
aemia, with partial pressure of oxygen in arterial blood
(PaO2) values of approximately 60 mmHg (Table 1b). In
Patient C, PaO2 gradually improved to 79.8 mmHg by
day 5. Respiratory function was assessed on Patient C
and initially revealed a slightly decreased diffusion capa-
city of the lung for carbon monoxide (DLCO) value of
13.36 ml min21 mmHg21. The value had returned to normal
when the test was repeated 3 months later. Respiratory
function was not assessed on Patients A and B because of
the impracticality of doing so under reverse isolation.
Remarkably, in Patient C, CT of the chest peformed on
day 1 revealed a crescent-shaped subpleural thickening in
the dorsum of the lower lobes of the lung, and adjacent
reticular shadows in the lung parenchyma, together sug-
gesting interstitial oedema of the organ (Figure 3a, arrow).
In accordance with the improvement of PaO2, a CT
scan taken on day 6 showed remarkable recovery, with
only minimal reticular shadows (Figure 3b), which further
became almost insignificant on day 60 (Figure 3c). Similar

Table 1. (a) Blood counts and serum chemistry on admission (5.5 h after exposure) to the NIRS. Values marked with an asterisk
are of the following day of the accident. (b) Arterial blood gas analyses on the day of admission. For Patient C, chronological
change is also shown.
a)
Patient Neutrophil (6109 l21) Lymphocyte (6109 l21) Haemoglobin (g l21) Amylase (IU l21) Uric acid (mg l21)
A 24.3 414 165 2143* 112*
B 12.5 268 129 2454* 98*
C 12.2 795 175 1094* 62*

b)

Patient Day pHa PaCO2 (mmHg) PaO2 (mmHg) HCO3 (mmol l21) BE (mmol l21) O2Sat (%)
A 0 7.553 33.9 62.4 29.2 7.0 94.6
B 0 7.574 36.8 55.1 33.3 10.7 92.7
C 0 7.525 42.2 62.6 34.1 10.2 94.0
C 3 7.435 40.5 64.3 26.6 2.2 93.2
C 5 7.402 42.5 79.8 25.9 0.4 95.8
C 13 7.413 45.0 80.8 28.1 2.9 96.0

PaCO2, partial pressure of carbon dioxide in arterial blood; PaO2, partial pressure of oxygen in arterial blood; BE, base excess; O2Sat,
oxygen saturation of arterial blood.
a
Initial pH values show iatrogenic alkalosis, which was caused by an administration of sodium bicarbonate at the National Institute of
Radiological Sciences for the purpose of decorporating uranium that might have been inhaled.

248 The British Journal of Radiology, April 2003

Tokai-mura criticality accident
findings were observed initially in Patient A (Figure 3d).
Also, in Patient B, mild subpleural thickening was observed
(not shown).
Clinical course
On admission, Patients A and B were kept under reverse
isolation. Initial therapeutic measures that were taken
common for all three patients included strict maintenance
of body fluid balance with the infusion of electrolytes and
plasma under central venous pressure monitoring and
selective digestive tract decontamination with oral admin-
istration of kanamycin sulfate (3 g day21), polymyxin B
(3 000 000 units day21) and amphotericin B (400 mg day21).
On day 4, vancomycin was substituted (2 g day21) for
kanamycin. Prophylactic systemic administration of anti-
biotics (imipenem 1 g day21), antiviral (acyclovir 1 g day21)
and antifungal (fluconazole 200 mg day21) reagents was
also started on day 1. Below the clinical picture of each
patient is described separately.
Patient A
Granulocyte colony-stimulating factor (G-CSF, 100 mg)
was administered intravenously on the evening of day 1.
Shortly after it was infused, the patient complained of mild
dyspnea and systemic rash. The symptoms resolved after
inhaled oxygen concentration was increased to 50%.
Electrocardiogram was normal on day 1, but evolved
inverted T waves in leads II, III and aVF on day 2,
possibly suggesting damage to the inferior myocardium.
Bone marrow taps were taken from the sternum and iliac
crest on day 1 and showed marked hypocellularity in both
the erythroid and myeloid lineages (not shown). The
myelogram of the smear from the sternum was as follows:
myeloblast 1%; promyelocyte 1%; myelocyte 3.6%; meta-
myelocyte 4%; band 32.4%; segmented 54.4%; eosinophil
1.4%; monocyte 0.8%; lymphocyte 1%; and plasma cell
0.2%. Some cells had intranuclear vacuolations, which
have also been reported in previous accidents [5]. White
blood cells (WBC) of Patient A slightly increased on
day 2, then rapidly decreased and almost disappeared
by day 7 (Figure 2a). Later we were informed that he
had been warned of leukocytosis for at least 2 years.
Lymphocytes kept decreasing and disappeared on day 3
(Figure 2a). Platelets also decreased steeply, necessitating
platelet transfusion starting from day 5 (Figure 2a).
Haemoglobin was rather elevated initially (Table 1a), pos-
sibly reflecting haemoconcentration, but then decreased
rather steeply by day 7 (Figure 2a, lower right panel, dia-
mond) without documented bleeding. From our evaluation
of the severity of his disease based on his symptoms and
signs in comparison to the victims of previous accidents
and the preliminary results of dosimetry that we received
on day 2, the spontaneous recovery of the bone marrow of
Patient A was judged to be quite unlikely. Therefore on
day 2 it was decided to treat the patient with haemato-
poietic stem cell transplantation and to transfer him to the
University of Tokyo Hospital, which was one of our
collaborators. G-CSF was discontinued for several days.
While in the NIRS hospital, the patient’s facial oedema
slightly improved. However, he complained of painful
forearm swelling on the right side on day 2, which
The British Journal of Radiology, April 2003
subsequently became tenser. Although the patient con-
tinued to have watery diarrhoea and complained of diffuse
abdominal tenderness by palpation, he was apparently well
on days 1 and 2, following which his general condition
rapidly deteriorated. He received peripheral blood stem
cell transplantation on days 6 and 7. His progressive
hypoxaemia necessitated endotracheal intubation on day 10.
Although the graft took, the patient continued to have
respiratory failure, subsequently exhibited severe skin
lesion and gastrointestinal bleeding, and died of multiple
organ failure on day 82. The detail of his therapy has been
published elsewhere [14].
Patient B
G-CSF (100 mg, intravenous infusion) was started on
day 1. This patient also exhibited systemic rash after the
infusion, it was decided to administer the growth factor
through round the clock infusion from day 2 (500 mg day21).
Bone marrow aspirates from the sternum and iliac crest on
day 1 were markedly hypocellular with some intranuclear
vacuolations. The myelogram of the smear from the
sternum was as follows: promyelocyte 0.4%; myelocyte
2.2%; metamyelocyte 2.0%; band 32.4%; segmented 58%;
eosinophil 1.2%; lymphocyte 2.2%; and phagocyte 1.6%.
WBC slightly increased on day 2, almost stabilized on
day 3 and then rapidly decreased and almost disappeared
by day 7 (Figure 2a). Lymphocytes also rapidly decreased
and disappeared on day 7 (Figure 2a). Platelets and hae-
moglobin decreased rather gradually (Figure 2a). We
anticipated that Patient B, although obviously having
been less severely irradiated than Patient A, would suffer
at least a prolonged period of severe leukopenia. As the
patient may have exhibited severe skin injury and gast-
rointestinal lesion during that period, we reasoned that it
should be more beneficial than detrimental to support his
leukopenic period with haematopoietic stem cell trans-
plantation. He was transferred to the Institute of Medical
Science, University of Tokyo, which was another of our
collaborating hospitals, and received umbilical cord blood
transplantation on day 9. The graft initially took, and then
was gradually replaced by his own haematopoietic cells.
The patient also had oedema in the right forearm in the
first several days, and later evolved severe skin lesions
involving the large part of his body surface, in particular
the face and extremities. Although the skin lesions were
extensively treated with skin grafts, he later evolved
gastrointestinal bleeding and infectious complications,
which caused him to die of multiple organ failure on
day 210. The detail of his therapy has been published [15].
Patient C
Judging from Patient C’s symptoms and signs, as well as
the preliminary results of dosimetry, we expected his bone
marrow to recover spontaneously. Therefore, he remained
at the hospital of the NIRS and was treated without
haematopoietic stem cell transplantation. The bone mar-
row aspirates from the sternum and iliac crest on day 1
showed decreased erythroid series and well preserved
myeloid series. The myelogram of the smear from the
sternum was as follows: myeloblast 0.4%; promyelocyte
2.8%; myelocyte 5.2%; metamyelocyte 4.6%; band 17.3%;
249

segmented 32.6%; eosinophil 3.4%; monocyte 1.8%;
lymphocyte 17.2%; plasma cell 1%; phagocyte 0.4%; baso-
philic normoblast 1%; polychromatic normoblast 5%; and
orthochromatic normoblast 6.8%. Some morphologically
abnormal megakaryocytes were also seen (data not
shown). The patient’s WBC count returned to normal
on day 1, then increased in response to G-CSF, which was
started on the evening of day 2 (Figure 2a). Neutrophils
then started to decrease following a stairwise pattern, and
reached a nadir of 1.096109 l21 on day 20 (Figure 2b).
The patient was kept under reverse isolation while having
neutropenia. Following the recovery of the neutrophil
count, G-CSF was reduced and eventually discontinued on
day 28. The decrease in platelets was slower than that of
the other two patients (Figure 2a), but necessitated platelet
transfusion on days 17, 20 and 23 (Figure 2b). Platelets
made a gradual recovery thereafter. The number of
lymphocytes was lowest on day 2 and also made a slow
250
T Hirama, S Tanosaki, S Kandatsu et al
Figure 2. Haematologic data of the patients.
(a) Peripheral blood counts of Patients A, B
and C during the first 11 days are depicted
with diamonds, rectangles and triangles, res-
pectively. (b) Haematologic data of Patient C
during the whole admission period. The
dynamics of neutrophil (upper left), platelet
(middle left), lymphocyte (lower left), haemo-
globin (upper right), reticulocyte (middle
right) and serum iron (lower right) are shown.
recovery (Figure 2a). Haemoglobin slowly decreased
from above 150 g l21 to 102 g l21 without any evidence
of bleeding (Figure 2b). Serum iron concentration steeply
increased from 134 mg dl21 on day 0 to 235 mg dl21 on
day 1, presumably reflecting the halting of erythropoiesis
(Figure 2b). It then decreased abruptly in the second
month, coinciding with the recovery of reticulocytes, which
had decreased during the first week and had then
undergone two transient rises (Figure 2b). During admis-
sion the patient exhibited spotty epilation as well as
marked diminution in the growth of beard. In addition, he
had a localized painless defect of the oral mucosa without
his knowing, which was pointed out on day 19. The lack
of pain might be attributable to inefficient inflammation
because of the neutropenia that peaked on day 20. These
symptoms were presumed to have been caused by
irradiation and improved gradually. He is now being
followed up in the outpatient clinic of the NIRS.
The British Journal of Radiology, April 2003
Tokai-mura criticality accident
Discussion
This accident caused ARS in three patients, who were
heavily exposed to mixed flux of neutrons and c-rays. ARS
develops when a large part of one’s body is exposed to a
high dose of ionizing radiation and has two characteristics.
First, its clinical manifestation depends upon the absorbed
energy of ionizing radiation: with increasing energy, sym-
ptoms resulting from the damage to the haematopoietic,
gastrointestinal, cardiovascular or central nervous system
prevail. Second, the syndrome has distinct clinical phases;
exposure is followed by the initial prodromal phase with
nausea and vomiting, followed by the latent phase, during
which the patient is asymptomatic, and eventually the
symptomatic phase. In the literature, the latent phase has
been described to be absent in patients who were exposed
to more than 8 Gy [9]. However, Patient A, who obviously
exceeded this dose, was apparently well for a few days
except for the tenderness on the abdominal wall and diar-
rhoea, suggesting that the initial evolution of symptoms in
even a severely irradiated patient may be overcome by
managing the medical condition appropriately.
After exposure to ionizing radiation, the severity of
damage to a given organ and the degree of the resultant
organ-specific clinical manifestations should be dependent
upon the absorbed energy of ionizing radiation. Accordingly,
early clinical manifestations of a radiation accident victim
are quite valuable for predicting the degree of ARS and
thereby formulating a reasonable therapeutic strategy.
Extensive efforts have been made to formulate methods to
predict the severity of ARS based on clinical manifestations,
in particular the dynamics of haematopoietic parameters [8,
11, 16], which outweigh other clinical parameters in its ease
of measuring, sensitivity and quantitativeness. Looking at
the blood counts of the present patients would therefore be
meaningful. The lymphocyte count of Patients A and B
decreased steeply in the first 48 h (Figure 2a). In Patient
The British Journal of Radiology, April 2003
Figure 3. CT image of the chest of Patients
A and C. Arrows show the relevant areas
stated below. (a) Day 1, Patient C. Note
bilateral subpleural thickening, more severe
in the right side, in the dorsum of the lower
lobes of the lung. Adjacent to the thickening,
restiform or rough reticular shadow can be
seen. These findings were also observed in
the dorsum of the upper lobes (not shown).
(b) Day 6, Patient C. Remarkable improve-
ment of the initial findings is obvious. (c)
Day 60 of Patient C. Only equivocal reticular
shadow remains. (d) Day 1, Patient A. Note
bilateral crescent form shadows. The density
of the area is higher than that of water, indi-
cating proteinacious exudate or soft tissue
thickening.
A, its rate of decrease was quicker than in Patient B, and
so extreme that it might be best comparable with that of a
victim of the Los Alamos accident in 1946 [5] who died
9 days after exposure. In contrast, its decrease in Patient C
was relatively mild. Thus the degrees of lymphocyte
decrease correlated well with the order of severity of ARS
among the three patients, and was compatible with our
initial reasoning that Patient A had received supralethal
dose of radiation. In addition, the predicted doses obtained
in the first 24 h relying on the graph by Baranov et al, which
were more than 10 GyEq, between 6 and 10 GyEq
and between 1 and 4.5 GyEq for Patients A, B and C,
respectively, seemed to be reasonable and were quite
valuable.
The dynamics of WBC also differed significantly among-
st the patients (Figure 2a). Although initial leukocytosis,
which reflected granulocytosis, was obvious in all of them,
WBC of Patients A and B further increased on the
morning of day 1, whereas that of Patient C returned to
normal. In addition, its increase in response to G-CSF was
only mild in Patients A and B (day 2), while that of
Patient C was quite significant (day 3). In Patients A and
B, WBC then decreased rapidly and almost disappeared on
day 7, whereas in Patient C it was favourably maintained
for another 10 days. Although the dynamics of the WBC
count differed between Patients A and B, we have to take
into account some conditions. (1) Patient A had been
warned of granulocytosis before the accident, which had
not been evaluated as to its exact pathology. His leuko-
cytosis after the accident, which was obviously more severe
than that of the other two, might partly reflect his higher
baseline reservoir of granulocytes. (2) In patient A, the
administration of G-CSF was discontinued on day 2 and
restarted on day 7, whereas in Patient B it was continued.
The quicker decline of WBC in Patient A on days 3 to 7,
compared with that in Patient B in the same period, might
reflect the lack of G-CSF stimuli. Due to these two
251

limitations, we would not consider the difference of the
WBC dynamics between Patients A and B significant.
Platelets decreased rather slowly when compared with
neutrophils or lymphocytes (Figure 2a). However, if we
look at the first 7 days, its rate of decline clearly separated
the three patients. For instance, platelet count of Patient A
turned below 50 G l21 on day 5, which again was
comparable with case 3 of 1946 Los Alamos accident. In
Patient B, the platelet count turned below 50 G l21 on day
8, which might be comparable with case V of the 1958
Vinca accident [17], who was the most severely irradiated
among the six victims in the accident and survived for
1 month. However, in the previous patient, the decrease in
lymphocytes and granulocytes were obviously milder than
those in Patient B. In Patient C the decrease was even
milder and fell below 50 G l21 on day 16. In this regard,
this patient might be comparable with case 1 of the 1945
Los Alamos accident who died 24 days after irradiation.
Because the use of G-CSF or other growth factors that
mobilize myeloid cells is now an important therapeutic
option in the treatment of ARS, the platelet count, which
is not supposed to be influenced by such reagents, might
make a reliable indicator of the severity of the disease
after several days. Interestingly but quite reasonably, the
dynamics of reticulocytes in Patient C showed a precisely
flipped pattern of that of serum iron. It started to recover
in the second week and exhibited two humps of transient
increase before its full recovery in the second month. The
transient rises of the reticulocyte count are seemingly
analogous to those of his neutrophil count, although they
came several days after those of the neutrophil count.
They might together represent limited proliferations of
damaged progenitor cells in the respective lineages.
During the initial period, the three patients exhibited
hypoxaemia. One explanation for this symptom would be
because ionizing radiation generates a certain amount of
ozone in the human body, which reportedly transforms
haemoglobin into methaemoglobin in vitro [18]. An
increased level of methaemoglobin in the blood would
then decrease PaO2. However, this is not likely to be the
case as studies have shown exposure of intact red cells or
rabbits to ozone did not result in a significant increase of
methaemoglobin [19, 20]. Another explanation would be
based on radiation damage to the lungs themselves.
Arterial blood gas analyses in the three patients showed
increased alveolar–arterial gradients of oxygen tension,
suggesting inefficient gas exchange (calculated results not
shown). In addition, respiratory function test in Patient
C revealed a transient decrease of the diffusion capacity
of the lungs. CT of the chest performed on day 1 showed
interstitial infiltration accompanied by subpleural thicken-
ing in two of them. In Patient C, the hypoxaemia gradually
improved, together with the CT findings. These findings
indicate the emergence of interstitial oedema of the lungs
caused by the damage to the endothelial cells in the organ.
Lung oedema has been observed in a recent fatal criticality
accident in Sarov 2 days after exposure [6] and in the Los
Alamos accidents in the post-mortem investigation [5].
Hypoxaemia shortly after a radiation accident has a
precedent in the Soreq accident in 1990 [21]. However, this
is the first time that the characteristic CT findings of the
chest and hypoxaemia in the very early phase of ARS are
described in a patient who has received a non-lethal dose
of radiation that does not cause any prodromal symptoms
252
T Hirama, S Tanosaki, S Kandatsu et al
of ARS. Hypoxaemia and the CT findings of interstitial
oedema of the lungs might make valuable early clinical
indicators of acute high dose irradiation.
The loss of consciousness in Patient A can be explained
as early transient incapacitation (ETI), which has been
observed in animals after whole body exposure to a very
high dose of radiation. According to Franz, ETI in
monkeys evolves 3–8 min after a whole body irradiation
and lasts for 5–20 min [22]. It is associated with a sudden
decrease in the cerebral blood flow and systemic hypoten-
sion [23, 24]. The cause of the symptom has not been
elucidated, but might reflect the response of the neuro-
vascular system [8] and might be related to a rapid release
of histamine [25]. MRI of the head of the patients on day 1
did not show any signs of cerebral oedema or focal lesions
(data not shown).
Elevation of serum uric acid has been described in two
lethally irradiated victims of the criticality accidents in
1946 and 1958 [3, 5]. Although we do not know the
mechanisms that underlie this phenomenon, we speculate
that it reflects non-specific and massive cell death caused
by high dose irradiation. In addition, increases in serum
amylase have been encountered after therapeutic irradia-
tion of the head and neck region and planned irradiation
of the whole body before haematopoietic stem cell trans-
plantation. The increase in amylase was mainly of salivary
origin, as has been described [26, 27]. As to its degree,
Patient B marked the highest level of serum amylase
amongst the three patients on day 1 after the accident,
followed by Patients A and C. Although the estimated
average dose in Patient B was lower than in Patient A, the
head of Patient B might have been very close to the source
of radiation (Figure 1). Thus, the extent of the amylase
elevation might have roughly reflected the severity of the
exposure to the head and neck in the three patients. Serum
uric acid and amylase should be measured in cases of
suspected exposure to ionizing radiation.
Conclusion
Our experience presented in this report is an important
example of applying state of the art therapy and extremely
intensive care to patients of ARS. It confirmed the value of
the early clinical signs as well as haematologic parameters
for predicting the severity of the disease. We have learned
that, even in a severe form of ARS, the initial hypotensive
period can be properly managed and that the damage to the
haematopoietic system may no longer be the direct cause of
death. Because a severe form of ARS involves multiple organ
systems, the rescuing of the bone marrow would subse-
quently confront us with the failures of other organ systems
such as the gastrointestinal tract, skin or lungs. Successful
treatment of the failures of such organ systems in an ARS
patient awaits further progress in the relevant fields at
the moment and will continue to challenge transplantation
medicine in the future. Finally, hypoxaemia and interstitial
oedema of the lungs may be considered an early indicator
of ARS, not only in severe cases but also in otherwise
asymptomatic cases.
Acknowledgments
We thank Dr Misao Hachiya for her general support,
and Ms Rika Hara for her secretarial help. We also thank
The British Journal of Radiology, April 2003
Tokai-mura criticality accident
Drs Hiroyuki Watanabe, Yoshihiro Yamaguchi, Toru
Iseki and Hideharu Tanaka and Professor Kazuhiko
Maekawa for generously providing us with their expertise
in the treatment of the three patients, and the Dose
Estimation Working Group for Three Victims, National
Institute of Radiological Sciences for sharing the results of
dosimetry.
References
1. Smith DR. Nuclear criticality accidents. Occup Med
1991;6:569–79.
2. Vargo GJ. A brief history of nuclear criticality accidents in
Russia—1953-1997. Health Phys 1999;77:505–11.
3. Lushbaugh CC. Acute radiation death resulting from an
accidental nuclear critical excursion: Clinical course of Case
K. J Occup Med 1961;(Suppl.):150–4.
4. Karas JS, Stanbury JB. Fatal radiation syndrome from
an accidental nuclear excursion. N Engl J Med 1965;272:
755–61.
5. Hempelmann LH, Lisco H, Hoffman JG. The acute radiation
syndrome: a study of nine cases and a review of the problem.
Ann Intern Med 1952;36:279–510.
6. International Atomic Energy Agency. The criticality accident
in Sarov. Vienna, Austria: International Atomic Energy
Agency, 2001.
7. Muramatsu Y, Noda Y, Yonehara H, Ishigure N, Yoshida S,
Yukawa M, et al. Determination of radionuclides produced
by neutrons in heavily exposed workers of the JCO criticality
accident in Tokai-mura for estimating an individual’s neutron
fluence. J Radiat Res (Tokyo) 2001;42(Suppl.):S117–28.
8. Fliedner TM, Friesecke I, Beyrer K, editors. Medical
management of radiation accidents—manual on the acute
radiation syndrome. London: British Institute of Radiology,
2001.
9. International Atomic Energy Agency. Diagnosis and
Treatment of Radiation Injuries. Vienna, Austria:
International Atomic Energy Agency, 1998.
10. Akashi M, Hirama T, Tanosaki S, Kuroiwa N, Nakagawa K,
Tsuji H, et al. Initial symptoms of acute radiation syndrome
in the JCO criticality accident in Tokai-mura. J Radiat Res
(Tokyo) 2001;42(Suppl.):S157–66.
11. Baranov AE, Guskova AK, Nadejina NM, Nugis V.
Chernobyl experience: biological indicators of exposure to
ionizing radiation. Stem Cells 1995;13(Suppl. 1):69–77.
12. Ishigure N, Endo A, Yamaguchi Y, Kawachi K. Calculation
of the absorbed dose for the overexposed patients at the JCO
criticality accident in Tokai-mura. J Radiat Res (Tokyo)
2001;42(Suppl.):S137–48.
The British Journal of Radiology, April 2003
13. Hayata I, Kanda R, Minamihisamatsu M, Furukawa M,
Sasaki MS. Cytogenetical dose estimation for 3 severely
exposed patients in the JCO criticality accident in Tokai-
mura. J Radiat Res (Tokyo) 2001;42(Suppl.):S149–55.
14. Ishii T, Futami S, Nishida M, Suzuki T, Sakamoto T,
Suzuki N, et al. Brief note and evaluation of acute radiation
syndrome and treatment of a Tokai-mura criticality accident
patient. J Radiat Res (Tokyo) 2001;42(Suppl.):S167–82.
15. Nagayama H, Misawa K, Tanaka H, Ooi J, Iseki T, Tojo A,
et al. Transient hematopoietic stem cell rescue using umbilical
cord blood for a lethally irradiated nuclear accident victim.
Bone Marrow Transplant 2002;29:197–204.
16. Andrews GA. Medical management of accidental total-body
irradiation. In: Hubner KF, Fry FA, editors. The medical
basis of radiation accident preparedness. Amsterdam, The
Netherlands and New York, NY: Elsevier, 1980:297–310.
17. Jammet H, Mathe G, Pendic B, Duplan J-F, Maupin B,
Latarjet R, et al. Etude de six cas d’irradiation totale aigue
accidentelle. Rev Fr Etudes Clin Biol 1959;4:210–25. [In
French.]
18. Gonet B. Properties of haemoglobin in ozonized haemoly-
sates of erythrocytes. Acta Haematol Pol 1989;20:167–75.
19. Freeman BA, Mudd JB. Reaction of ozone with sulfhydryls
of human erythrocytes. Arch Biochem Biophys 1981;208:
212–20.
20. Ross BK, Hlastala MP, Frank R. Lack of ozone effects on
oxygen hemoglobin affinity. Arch Environ Health
1979;38:161–3.
21. The radiological accident in Soreq. Vienna, Austria:
International Atomic Energy Agency, 1993.
22. Franz CG. Effects of mixed neutron-gamma total-body
irradiation on physical activity performance of rhesus
monkeys. Radiat Res 1985;101:434–41.
23. Cockerham LG, Cerveny TJ, Hampton JD. Postradiation
regional cerebral blood flow in primates. Aviat Space Environ
Med 1986;57:578–82.
24. Cockerham LG, Hampton JD, Doyle TF. Dose dependent
radiation-induced hypotension in the canine. Life Sci
1986;39:1543–7.
25. Doyle TF, Curran CR, Turns JE, Strike TA. The prevention
of radiation-induced, early, transient incapacitation of
monkeys by an antihistamine. Proc Soc Exp Biol Med
1974;145:1018–24.
26. Barrett A, Jacobs A, Kohn J, Raymond J, Powles RL.
Changes in serum amylase and its isoenzymes after whole
body irradiation. BMJ 1982;285:170–1.
27. Junglee D, Katrak A, Mohiuddin J, Blacklock H, Prentice
HG, Dandona P. Salivary amylase and pancreatic enzymes in
serum after total body irradiation. Clin Chem 1986;32:
609–10.
253