4. Second Henri Jammet Memorial Lecture. Radiation Accidents - an Overview and Feedback, 1950 - 2000 - Nénot, J-C.

Nénot, J-C.
Institut de Protection et de Sûreté Nucléaire, France

A. INTRODUCTION

Accidents that involve ionising radiation are rare. The damage they cause, although generally non-specific, has particular features, which can complicate the situation. For example, acute radiation syndrome has a latency period, which makes it impossible to predict the critical phase and may lead to unjustified optimism. While direct effect impact on those who have actually been exposed, some indirect effects may impact on large groups who have not been subjected to accident exposure. These indirect effects have been particularly well described in certain large-scale accidents, e.g. Chernobyl. Depending on their type and the circumstances in which they occur, radiation accidents assume very different forms. The range of accidents is extremely wide and varied. In the simplest cases, there is only one person injured, the source is known, diagnosis is easy, and the medical input is confined to discussion between physician and victim, thus corresponding to the features of normal medical management. The situation is often more complicated, either because of medical problems (doubtful diagnosis, uncertain prognosis and difficulty or disagreement in choice of treatment), or management problems (large numbers of victims, combined injuries, which complicate the choice of the most appropriate management), or finally because of the scale of the accident.

Out of these general considerations, a classification of accidents emerges that, like any classification, is arbitrary, but that above all takes account of the difficulties involved in the management. Management will depend on the way accidents come to light, their extent, the medical and/or public health problems they present, etc. This approach allows a distinction between accidents that present no special problems, those that are medically difficult to handle and those that bring major resources into play. A separate category includes accidents that are kept secret, generally for military or political reasons. In each group, there are accidents that are diagnosed from the start and those that are only discovered later. Examples chosen to illustrate this classification are given in Table 1, which is in no way intended to be exhaustive.

Table 1: Classification of radiological accidents, in terms of management difficulties

The number of accidents, that are only diagnosed when the symptoms presented by the victims suggest the radiological nature of the accident, seems to be increasing with time. Indeed, it often happens that the accident only comes to mind after suspicious deaths have alerted the authorities, and the accident is therefore only discovered by chance. In view of this, it is legitimate to wonder about the existence of serious, as yet unrecognised, accidents, whose victims have had their deaths attributed to «normal» causes.

B. ACCIDENTS THAT ARE RECOGNISED IMMEDIATELY

B.1. Easy to handle

The common aspects that make these accidents easy to handle fall into various categories: awareness of the accident is immediate and possible victims are unambiguously identified. The medical problems are analogous to those found in common disorders and do not exceed available logistical resources. However, mistakes of all kinds may occur, particularly with regard to diagnosis and treatment. The examples are drawn from three accidents in industrial radiation facilities and from a criticality accident in a military facility, all of which caused the death of a worker. The first three accidents involved whole body irradiation at supralethal doses when the worker walked into the irradiation room where the cobalt 60 source was in irradiation position. The fourth accident was caused by human error, which was possible because of defaults at the conception of the experimental device designed for the study of fission reactions.

In Kjeller, Norway (2 September 1982), it was estimated that exposure was greater than 10 Gy [1]. The worker died thirteen days after the accident of acute kidney failure, with underlying marrow aplasia.

In Soreq, Israel (21 June 1990), the victim received high quality care; he was hospitalised two hours after the accident. The absorbed dose was evaluated between 10 and 20 Gy. He started growth factor therapy on the first day [2]. Bone marrow transplantation was carried out on the fourth day using his brother's marrow. The engraftment was confirmed. The patient died 36 days after the accident from severe intestinal syndrome and respiratory failure.

In Nesvizh, Belarus (26 October 1991), the severity of the irradiation was immediately recognised, and the victim was very rapidly transported to Moscow [3]. The dose was estimated between 9 and 16 Gy. The physicians believed that re-population from less exposed areas was possible and that the risks of marrow transplantation outweighed its advantages. The patient was given growth factor therapy. Haemopoiesis was restored in a few weeks, but there was a rapid and dramatic deterioration of all body systems. A gastrointestinal syndrome appeared after a week, followed by extensive cutaneous involvement. Liver and kidney failure ensued and was finally complicated by pulmonary problems. The patient died of respiratory problems 113 days after the exposure.

In Sarov, Russia (17 June 1997), in the military centre known as Arzamas-16, a number of factors contributed to the criticality excursion which occurred during a routine manipulation; the operator was exposed to neutrons with little gamma contribution to total dose [4]. The mean dose was evaluated at around 15 Gy, with a very non-uniform distribution within the body, of which some areas had received up to 60 Gy. The victim died of an acute heart insufficiency, less than three days after exposure. These apparently simple cases provided valuable lessons:

As well as the dose level, the distribution of exposure to the body is determinant of therapeutic decisions. Bone marrow transplantation is indicated only in the case of homogeneous exposure at high level. This statement underlines the need for the physicians to rely on physical data, such as those obtained from the reconstruction of the accident.

The bone marrow should not be seen as an isolated organ, since irradiation induces a series of interdependent health effects, which occur at a rate that is dependent on cellular dynamics. Particular attention should be given to the effects of radiation in several other organs.

The Soreq and Nesvizh accidents were the first occasions where growth factors were used from the start of the conditions. It is difficult to assess their effectiveness, given the transplantations performed in one accident and the repeated transfusions in the other. However, it was concluded that growth factor therapy contributed to the rapid engraftment and cell maturation in the first accident, and to a faster haematological recovery in the second.

B.2. Difficult to handle

The management of a radiation accident becomes difficult as soon as several people are involved. In general, the difficulties are not of a medical nature, although the structures required to cope can easily be overwhelmed when there are a number of actual or possible victims. Handling is often complicated by the fact that group accidents attract media attention that is not always justified. One accident due to a gammagraphy source and a criticality accident, which both took place in 1999, provide examples of the pitfalls inherent in these events.

In Yanango, Peru (20 February 1999), a 37 year old welder picked up iridium-192 source of about 1 TBq that he had found on the ground and placed it in the right back pocket of his trousers; he continued the work for 6 hours and then went home. The plant staff found the source in his home 10 hours after the beginning of exposure, which involved the whole family (wife and three young children) during some hours. The severity of the induced lesions of the tight increased rapidly. In May, Peru requested the assistance of France where modern treatment resulted in a temporary improvement but could not avoid a dramatic relapse after six months leading to amputation of his right leg that neither the clinical evolution nor the dosimetric findings forecasted. One year after the accident the patient was in a hopeless situation.

In Tokai Mura, Japan (30 September 1999) three workers were involved in a criticality accident, which happened at a uranium conversion facility. A solution of enriched uranium amounting to 16 kg which was several times more than the specified mass limit (2 kg) and the critical mass (5.5 kg) had been poured directly into a precipitation tank, bypassing a dissolution tank and buffer column intended to avoid criticality. This action is assumed to have been done in order to earn time. The ratio neutron/gamma was about 60/40. The medical management was performed with the most reliable and modern means; two victims were in severe aplasia complicated by large and deep radiation burns. The most severely exposed patient (about 20 Gy) received a haematopoietic growth factor (G-CSF) and a stem cell transplantation, which improved the haematological state; this patient died 83 days after exposure from multiple organ failure (kidney, liver, heart and lung). The second patient (about 8 Gy) received cord blood transplantation, of limited efficiency, since his own marrow started its spontaneous repair within a few weeks [7]. His death on 211^st day was also due to multiple organ failure. The lessons, which can be derived, are the following:

This accident demonstrates that the human factor plays a major role in the genesis of accidents. Grave errors of judgement, leading to decisions with direct consequences, seem astonishing in qualified and experienced technicians.

Dosimetric reconstruction, when based on reliable information, provides irreplaceable elements for decision making and avoids a lot of suffering. However, it cannot predict the clinical evolution, in particular long term relapses in vital organs with slow evolution injuries.

The efficiency of new therapeutic means should hide the fact that a whole body irradiated patient suffers from multiple injuries and that haematopoetic cell transplantation, whatever their origin and the importance of side effects, has limited indications in radiation accidents.

Exposures with a substantial neutron component result in specific injuries, which are difficult to handle.

B.3. Catastrophic accidents

An example of an immediately recognised catastrophe is the Chernobyl accident, Ukraine, (26 April 1986) which, on its own, brought together a large number of direct and indirect effects of radiation and affected the whole of the Northern hemisphere. The final count of people experiencing acute effects was 31 dead, 28 of them from radiological causes. Of the 56 patients with cutaneous radiation syndrome, regardless of the haematological treatment (13 bone marrow and 6 foetal liver grafts with only two survivals) 14 died primarily because of their cutaneous lesions [9, 10]. Of the several hundred-thousand workers who for months worked in shifts on the decontamination of the site, a large, but unknown, number received significant doses. Within the population, while no excess of leukemias has been demonstrated, there is an unquestionably high rate of childhood thyroid cancer in the regions of Ukraine, Belarus and Russia, that were particularly affected by fallout in the early days. In 1996, the number of childhood thyroid cancers stood around 800 in 1996 [11]. The total number of thyroid cancers, which can be predicted, is estimated at about 10,000, but it is impossible to assess a mean likely value [12, 13]. In 1999, the number of childhood thyroid cancers in Ukraine, Belarus and Russia is estimated at around 1,500, but it is not possible to predict whether this excess will continue, become stable or decrease [14, 15]. From the date of this accident, the factors that concern handling are:

The expectations of bone marrow transplantation were not fulfilled. It was concluded that the indications for bone marrow transplantation in the treatment of accidental irradiation are limited and that the range of doses for which transplantation may be envisaged is higher than 8 Gy.

Medical handling when there is a large number of victims (237 diagnoses of acute radiation syndrome were made initially at Chernobyl) is very difficult. There is no doubt that, without meticulous preparation and emergency plans based on potential scenarios, the situation will get out of control and end in fiasco.

The same general observations apply to protecting the population. Providing shelter and stable iodine to the population as well as carrying out temporary or permanent evacuation raises difficult logistical, socio-psychological and economic problems.

The emergence of an excess of childhood thyroid cancers has raised a polemic, which would not have occurred if the necessary monitoring had been set up at the right time. The resources required to treat the consequent cases are considerable and a sustained commitment will be needed.

C. ACCIDENTS WITH DELAYED RECOGNITION

The number of accidents for which the radiological aetiology is recognised by chance with considerable delay from the beginning of the exposure is increasing with time. Often the type of injuries allows the physician to evoke the origin of the accident. The question of the number of severe accidents, which remain completely unknown and are attributed to common causes cannot be exactly answered.

C.1. Limited number of victims

There is a number of examples of such events, due to either industrial or medical sources. The following examples provide a relatively good overview of accidents, which occurred in the industrial (six accidents), and medical (one accident) fields.

In Mexico City (21 March 1962), a whole family was wiped out by a cobalt-60 source, found on a dump [16, 17]. A 10 year old boy died in April, then his mother in July, without the cause of death being identified. It was only on the death of a girl of 3 in August that the common origin of these deaths was discovered. Nonetheless, there was a further unfortunate death in October. The father's survival was due to the shorter periods he spent in the house. In Brescia, Italy (13 May 1975), a worker in an industrial cereals irradiation facility with four cobalt-60 sources, entered the irradiation room by climbing onto the conveyor belt. His symptoms were attributed to the toxic effects of insecticides. For more than two days, his exposure to one of the sources, which was not shielded when he entered the room, remained unknown to the physicians at the hospital where he was admitted. After his hospitalisation in a specialised unit, the average dose to the bone marrow was evaluated at 12 Gy; with a non-uniform distribution. The victim died after thirteen days [18].

In Setif, Algeria (5 May 1978) and in Casablanca, Morocco (19 March 1984), iridium-192 sources were lost and picked up by families. The diagnoses were only made after 38 and 80 days, respectively [19, 20]. The accidents were only revealed by chance, after a physician identified the aetiology of the symptoms. In all, the Algerian accident resulted in the death of a 47 year old woman, in four serious life-threatening whole body overexposures affecting young women aged between 14 and 20, and multifocal localised overexposures in two children aged 3 and 7. The Moroccan accident resulted in eight deaths; four young children died from severe aplasia a few days after their parents. Several relatives were affected by bone marrow depression at various degrees.

In San Salvador, El Salvador (5 February 1989), three workers at an industrial sterilization plant using a cobalt-60 source were irradiated when attempting to unblock the source holder [21]. The irradiation was only identified when a burn appeared on the third day. It was only three days later that the accident was diagnosed, and it was two weeks before the irradiator was finally rendered safe. The doses received by the three main victims were evaluated as 8, 4 and 3 Gy, with areas of the body exceeding 10 Gy in the most exposed victim. Two patients each had a leg amputated. One had a second leg amputated and the other developed a respiratory disorder; this was the indirect cause of his subsequent death during an operation.

Following a malfunction of the linear accelerator used by Zaragoza Hospital, Spain (10 to 12 December 1990), 27 patients were exposed to higher than intended doses [22]. The cause of a breakdown was wrongly identified and the subsequent repair resulted in a change in the energy of the electrons. This was the primary cause of the overdose. Patients developed pulmonary, oropharyngeal and bone marrow lesions, complicated by vascular and skin damage. It was estimated that at least twelve patients died as a result.

In Forbach, France (13 August 1991), three handlers in a linear accelerator unit were irradiated during repeated operations on the machine, in which only the electron source was cut off, but accelerator voltage maintained to save time [23]. In these conditions, the residual dose rate was about a few Gy per second. Initially, the skin lesions were attributed to the sun. It was only after several days that the cause was identified, as a result of serious deterioration in one of the victims, who was hospitalised. This patient underwent repeated skin grafts and remained in hospital for a whole year. In Hanoi, Vietnam (17 November 1992), an engineer suffered severe hand irradiation from a physics institute linear accelerator beam, while positioning a sample for analysis [24]. Although the victim was immediately aware of the accidental exposure and reported it, his burns were only linked with this event two weeks later when his clinical state became serious. Four months later, the victim was transferred to a specialised hospital in France where he stayed for more than one year. He had to have his right hand and the last two fingers of his left hand amputated.

The lessons of these accidents are varied in nature:

Too many small sources get lost. These "orphan" sources often result in severe injuries to large numbers of people, who often obtain medical help too late. Absence or inadequate regulation are aggravating factors.

In large industrial plants, accidents occur through a lack of elementary safety rules, through poorly trained staff who are not warned of the risks, and through failure to observe rules and instructions, where these exist.

The hospital accidents that become public knowledge involve overdosages. It should not be forgotten that underdosages, which do not produce immediate effects, could also have serious consequences for patients. However, it should be recognised that accidents in hospitals are rare as a proportion of all treatments.

C.2. Large number of victims

Certain accidents, that are not recognised until some time has passed, result in large numbers of people being exposed at very varied levels during the movements of radiation sources in space and time. The probability for large groups of individuals to be exposed increases with the time during which the source is out of control. Usually, the sources are lost ones, but occasionally accidental medical irradiation may be involved. Four accidents that occurred in hospitals and three cases of lost sources furnish illustrations of this type of accidents.

The first serious accident with a medical source happened in a hospital in Columbus, Ohio, USA. Between 1974 and 1976, a wrong calibration, due to an error in the cobalt 60 half-life, caused the overexposure of 426 patients; their doses were 15-45% higher than the prescribed doses, depending on the time where they received their treatment [25]. Among the 183 patients who were still alive one year after their treatment, more than one third had severe complications of the central nervous system (brain and spine) and the gastrointestinal tract (oropharynx, colon, rectum). An opposite result was discovered at the hospital of Stokes-Upon-Trent (UK) after the underexposure of 1045 patients, who received, between autumn 1982 and winter 1991, doses lower than the prescribed ones, because of an error in the treatment procedures [26]. The consequences of this underexposure, 5-35% less than expected, remain imprecise. The third accident occurred between June 1985 and January 1987, in four hospitals in the USA; the same mistake, on the same type of machine, a THERAC 25, caused five series of overdose accidents [27]. The initial event was an operator error in programming the machine, an error that was actually indicated on the checking report. The last accident occurred despite all the hospitals using this type of machine having been warned, and a safety procedure laid down. This series of accidents caused severe tissue damage, including burns, myelitis, paralysis and complications that resulted in deaths. The fourth accident happened in Indiana, Pennsylvania, USA (16 November 1992); an elderly patient returned to the post-treatment unit without the brachytherapy indium source (high dose rate) being removed [28]. Four days later, a nurse threw a catheter discharged by the patient - and containing the source - into the waste. The next day, the patient died without her death being attributed to irradiation. During transport of the waste more than 90 people were exposed; fortunately their doses were not high enough to produce acute effects. A monitoring device located at the entrance to a waste processing facility finally raised the alarm.

The first example of the loss of an industrial source is provided by the accident in Xinzhou, Shenzi, China (19 November 1992). The cobalt sources of an industrial facility, closed in 1980 after twenty years in operation had been stored under six metres of water in a well [29]. When a building site was opened, a worker found one of these sources with a residual activity of 400 GBq, and took it home. He died in hospital fifteen days later, followed by his father and his brother. As the patient had come to the hospital with the source, nursing and medical staff and visitors were exposed, together with the workers who transported it to the waste storage site. It was only after 2-3 weeks that the real cause of the problem was discovered. Among the twelve most exposed individuals, the doses ranged between 1.2 and 2.3 Gy. The source was only found 76 days after it was taken. The second accident occurred in Tammiku, Estonia (21 October 1994), after a caesium 137 source was stolen from a storage centre and kept in a house for 27 days. In this period, the death of a young man was put down to traumatic toxaemia. It was not until a 14 year old boy was found to be suffering from a haematological disorder and radiation burns that the aetiology was established. In all, seven people were exposed to significant doses, and five suffered of serious damage [30, 31]. More recently, eleven young frontier guards were exposed to a high activity caesium-137 source at the Lilo military training centre, in Georgia. The victims were exposed for approximately one year, from mid-1996 to April 1997. The accident was recognised only at the end of August 1997 [32]. Each of the victims suffered from one or more localised radiation injuries. The four most severely injured victims, all of them with multiple burns, were treated in France, while seven others were hospitalised in Germany. Some patients were treated with success for the first time in the case of radiation burns, with new techniques, such as artificial skin grafts. Several individuals other than these victims might have been exposed, since over 200 further sources of low activity were discovered by systemic grid scan, in the training site.

These accidents, each very different in its origin (medical, storage of materials, research institute) and their consequences, have a common feature: radiation protection measures based on common sense could have prevented deaths and serious injuries.

The number of accidents caused by military sources is increasing. The lack of instructions and regulations at army departure, combined with disinterest of military responsible officers give an odious character to this type of accident.

C.3. Severe consequences for the population and environment

Two accidents with very heavy health consequences had in addition severe impact on the environment and caused extensive socio-economical damages. Both of them were caused by an abandoned radiotherapy source with dissemination of radioactive materials. Another accident, involving more than one hundred patients treated in a hospital, was a real national drama.

In Juarez, Mexico (6 December 1983) a teletherapy source consisting of pellets of cobalt 60 with a total activity of 15.6 TBq was dismantled without subsequent control of disposal. The problem was discovered by chance six weeks later when concrete reinforcing irons triggered alarms in the Los Alamos centre in the USA [33]. These irons came from a foundry, which had processed residues recovered from the source. It was estimated that 4000 people had been exposed. Of these, about 800 received doses above 50 mGy, eight between 1 and 7 Gy, delivered over a two month period. No deaths resulted, since exposure was spread over time. The Goiânia accident, Brazil (13 September 1987) was more severe than the Juarez, as the teletherapy source consisted of highly soluble caesium-137 chloride. The first signs of acute radiation syndrome were attributed to a tropical disorder, and two weeks elapsed before the accident was recognised [IAEA 1988 and 1998]. Twenty people required emergency hospitalisation, ten received doses between 3 and 7 Gy. Most of the victims were suffering from whole body irradiation and radiation burns, and many had internal contamination some resulting in high committed doses. Four victims died and several had to undergo surgery, some of which was radical. Initial decontamination of the environment, involving the demolition of houses, took place after only three weeks. The town and its surroundings were not considered acceptably free of contamination until March 1988. The total economic impact was very severe, and the whole regional economy was affected.

The accidental overexposure of 115 patients, including children, in San Jose, Costa Rica, between 26 August and 27 September 1996, was considered a national tragedy. Errors in the calibration of the cobalt 60 teletherapy unit resulted in overexposures of about 50-60 % more than the prescribed dose [35]. This overexposure resulted in severe health consequences, which were in addition to the illnesses for what the patients were treated and in several cases were worsened by the treatment procedures (insufficient fractionation, maladjusted fields, etc.). This accident caused dramatic effects in 4 patients (quadriplegia, paraplegia, spinal cord demyelinisation, severe digestive and cutaneous disorders), marked effects with might worsen in 16 patients, some risks for the future in 26 patients; only 22 patients did not suffer from the accident, mainly because treatments were interrupted before overdose was realized. The number of deaths directly due to the accident is difficult to assess [36]. Among the 61 deaths in two years, it is likely that 17 were caused by errors in calibration leading to complications, which might have caused their deaths. Among the 51 patients still alive in October 1998, two were suffering from catastrophic complications and twelve exhibited marked and invalidating sequelae.

The lessons from these accidents, all three related (directly or indirectly) to medical sources, are:

Study of the scenarios shows that there is an unacceptably high potential probability of these sorts of accidents occurring.

For prevention, there are factors in common between these catastrophic accidents and those resulting from the loss of small industrial sources: absence or lack of regulation, failure to control sources, inadequate training and awareness of staff.

For medical management, a common factor is the inadequacy of resources.

The Costa Rica accident underlines the importance of quality assurance and education of personnel for the prevention of errors which may result in severe consequences. In case of overexposure, it is essential to organize a follow-up system, which should include major medical and psychological support during at least five years; in addition a registry of data on patients should be set up. This will allow for improving the treatments, to comfort the patients and to derive knowledge on the effects of radiation at dose levels, which are not current either in practices, or in accidents.

D. ACCIDENTS THAT ARE KEPT SECRET

The secret accidents generally belong to the military arena, and were especially common during the Cold War. The improvement in east-west relations does not, however, allow us to affirm that complete openness has become the norm for accidents that may have political or military implications, or that could involve revelation of details considered to be secret. The great nuclear powers have perhaps not yet fully lifted the veil of secrecy over all the incidents and accidents that punctuated the arms race. The more typical examples come from the disastrous management of Soviet military installations in the fifties and sixties, and from the loss of nuclear devices at sea by both superpowers.

In Chelyabinsk, Urals (29 September 1957), a very large stretch of land in the Ural Mountains was contaminated by the accidental release of fission products from a secret installation, dating from the immediate post-war period [37]. This accident was only revealed in 1990. The resulting exposure of the population and the work force came on top of the exposure they were already experiencing from operational practices that were considered normal at the time. Following this accident, twenty villages of 7500 people had to be evacuated permanently. The most exposed groups of the population show a significant increase of leukaemia. The accident, which took place in the Atlantic in 1961, shows the potential consequences of the quest for secrecy at any price. To prevent a nuclear submarine being recovered by other powers, the Soviet authorities ordered the crew to carry out makeshift repairs. During these operations, several members of the crew received very high doses, and at least eight of them died as a result. Although losses at sea do not correspond to the definition of nuclear accidents, nuclear weapons lost at sea following air accidents, after failed nuclear missile launches, reactors of ships and especially submarines that have sunk, and plutonium based reactors from satellites that have broken up when re-entering the earth's atmosphere, should be mentioned. The American list for losses at sea, accounts for: seven incidents involving several nuclear weapons (1950-1965), three missile launches, two submarines lost with their reactors and three satellites [38]. On the Soviet side, most of the losses came from the submarine fleet: between 1968 and 1989, seven weapon armed submarines sunk, in addition to a destroyer lost in the Black Sea in 1974. The lessons of these accidents, with their varied causes and consequences, are evident:

Military imperatives put the accent much more on secrecy and the military objectives than on concerns for safety and protection. Only a radical change of mentality, initially at governmental level, underpinned by a commitment to openness, would enable military installations and their associated operations to attain safety levels comparable with the best civilian facilities.

E. CONCLUSIONS

The lessons, which may be learned from the feed-back, can be divided into two broad categories: (1) diagnosis of the accident and of the casualty, (2) technical management of the accident and treatment of the injured victims.

There are too many cases where it takes too long for the accident to be identified. It is conceivable that accidents occurring in similar circumstances have remained unidentified. Whatever the period that elapses between the event and its discovery, wasted time has grave consequences: increases in the number of casualties, deterioration in their condition; difficulties in diagnosis and treatment. Prevention is essential and must work at all levels: improvement or introduction of appropriate regulations, training of staff and facility managers, commitment to openness. Attention to these matters, which are simply common sense, would help reduce the number of sources lost. Special attention should be paid to the human factor, since the sequence of causal events in most of the accidents includes human errors involving serious violations of the most elementary safety rules; these need to be foreseen and prevented.

The degree of severity is not immediately obvious, based on clinical criteria alone. Physical dosimetry is not always accorded appropriate significance, since physicians may not have the necessary information and training. The day-to-day concerns of these physicians are very far from the specific problems presented by radiological accidents. Biological dosimetry is undoubtedly an essential stage in the estimation of the dose. However, only the combination of both the physical and biological methods provides the two parameters that are essential to diagnosis and prognosis: absorbed dose to the bone marrow and the distribution of exposure to the body. Only careful planning for these accidents can prevent serious errors of assessment.

Feed-back is particularly valuable. Handling the public health aspects of a large-scale accident involving large numbers of actual or potential victims requires considerable resources. The first condition for the effective handling of such crises is for the authorities to recognise the scale of the problem, and to have made continuous plans at every level. Experience shows that psycho-sociological factors can outweigh medical or even public health factors. However, the list would be incomplete if reference was not made to the provision of information. It must flow in all directions and bring into play all the networks: authorities, physicians, experts, casualties, the population and the media.

A radiation accident may assume very different forms and may not present the same medical and public health problems, and hence not involve the same specialists. Acute irradiation of the whole body requires full scale intensive care, in particular haematological treatment, while localised exposure will raise problems akin to those presented by burns, which may be deep or superficial, more or less extensive, and may need very precise surgical treatment. Internal contamination, whether or not combined with external contamination, requires the skills of particular specialists. The choice of treatment after serious haematological radiation injury has caused unjustified debate. It is true that a radiation casualty never fits into a stereotypical framework, but each patient is a specific case to be treated in a specific way. The treatment of aplasia has progressed considerably in the last twenty years. Patients may survive the radiation-induced aplasia phase, but be threatened by the failure of other organs or systems. Table 2 illustrates that, after severe overexposure, death occurred because of multiple organ failure. Only time will tell if the hopes invested in growth factor therapy will prove justified. Methods of successfully coping with other syndromes remain imperfect and research efforts in these areas must be maintained.

Medical management of localised overexposures, when skin basal layer receives doses which result into necrosis and when the size of burns is important, is very difficult. It should be recognised that little progress has been done during the last decades. However, the application of new techniques applied to conventional burns to radiation-induced lesions authorizes good hopes, for diagnosis and treatment. One of the victims of the Georgian accident, who presented a living illustration of various forms of radiation burns (33 different lesions), provides an example of progresses in new techniques. Computerised dosimetric reconstruction and magnetic resonance imaging made it possible for the first time to assess the extent of the injuries. The use of artificial skin demonstrated its advantages when compared to classical methods. Experience is still too limited to ascertain that these techniques will be the best treatment of such injuries, but there are nowadays no contradictory arguments.

Table 2: Lethal radiation accidents. Types of medical treatment and causes of death

* G.I.T.: gastro intestinal tract; GVHD: Graft-Versus-Host-Disease

** Victims of the Israel and Japan accidents, which appear in the «transplantation» category, received also growth factors (Israel: GM-CSF + IL-3, Japan: G-CSF).

REFERENCES

[1] Stavem P., Brogger A., Devik F., et al. Lethal acute gamma radiation accident at Kjeller, Norway, report of a case. Acta radiologica Oncology, 24, Fasc. 1, (1985) 61-63.

[2] International Atomic Energy Agency. The radiological accident in Soreq. IAEA, Vienna (1993).

[3] International Atomic Energy Agency. The radiological accident at the irradiation facility in Nesvizh, IAEA, Vienna (1996).

[4] International Atomic Energy Agency. The criticality accident in Sarov. IAEA, Vienna (2000).

[5] Nénot J.C. Accident grave d'iradiation au Pérou avec une source d'iridium. Revue Générale Nucléaire, n°3, mai-juin, (1999) 73-74.

[6] Institut de Protection et de Sûreté Nucléaire. Fiche relative à l'accident survenu le 30 septembre 1999 à Tokaï-Mura (Japon). 14 octobre 1999.

[7] International Atomic Energy Agency. Report on the preliminary fact finding mission following the accident at the nuclear fuel processing facility in Tokaïmura, Japan. IAEA, Vienna (1999).

[8] Gourmelon P. Personal communication (1999).

[9] Baranov A.E. and Guskova A.K. Acute radiation disease in Chernobyl accident victims. The medical basis for radiation accident preparedness II. Clinical experience and follow-up since 1979. Ricks R.C. and Fry S.A. Eds. (Elsevier/North Holland, New York, 1990) 79-87.

[10] Gale R.P. Immediate medical consequences of nuclear accidents: lessons from Chernobyl. Jama, Aug 7, 258, n°5, (1987) 625-628.

[11] Williams E.D. Effects on the thyroid in populations exposed to radiation as a result of the Chernobyl accident, One decade after Chernobyl. Summing up the consequences of the accident, IAEA, Vienna, (1996) 207-238.

[12] Cardis E. Estimated long term health effects of the Chernobyl accident, One decade after Chernobyl. Summing up the consequences of the accident. IAEA, Vienna, (1996) 241-279.

[13] Nénot J.C. The Chernobyl accident: the consequences in perspective, One decade after Chernobyl. Summing up the consequences of the accident. IAEA, Vienna, (1996) 475-509.

[14] Tronko M.D., Bogdanova T.I., Komissarenko I.V. et al., Thyroid carcinoma in children and adolescents in Ukraine. Cance, 86 (1), 1999 July 1., (1999) 149-156

[15] Ivanov V.K., Gorsky A.I., Tsyb A.F. et al. Dynamics of thyroid cancer incidence in Russia following the Chernobyl accident. J. Radiol. Prot., Vol 19 No 4, (1999) 305-318.

[16] Andrews, G.A., Mexican Co 60 radiation accident. Isotopes and radiation technology, vol 1, n° 2, (1963) 200-201.

[17] Martinez R.G., Cassab G.H., Ganem G.G. et al. Observations of the accidental exposure of a family to a source of cobalt-60. Revista Medica Inst Mex Següro Social, 3, (1964) 14-68.

[18] Jammet H., Gongora R., Pouillard P., et al. Clinical and biological comparison of two acute accidental irradiations: Mol (1965) and Brescia (1975). The medical basis for radiation accident preparedness. Hübner K.F. and Fry S.A. Eds. (Elsevier/North Holland, New York, 1979) 91-104.

[19] Jammet H., Gongora R., Pouillard P. et al. The 1978 Algerian accident: four cases of protracted whole-body irradiation. The medical basis for radiation accident preparedness. Hübner K.F. and Fry S.A. Eds. Elsevier/North Holland, New York, (1979) 113-129.

[20] Parmentier N., Nénot J.C., Parmentier C. Two cases of accidental protracted overexposure: aspect of an extensive bone marrow study, The medical basis for radiation accident preparedness II. Clinical experience and follow-up since 1979. Ricks R.C. and Fry S.A., Eds. (Elsever, New York, 1990) 29-51.

[21] International Atomic Energy Agency. The radiological accident in San Salvador, IAEA, Vienna, (1990).

[22] Esco R., Lopez P., Bellosta R. et al. Accidental overirradiation syndrome. Radiotherapy and oncology, 28, (1993) 177-178.

[23] Chanteur J. Un accident d'irradiation. Médecins et rayonnements ionisants. n°3, Juillet 1992, (1992) 5-6.

[24] International Atomic Energy Agency. An electron accelerator accident in Hanoi, Vietnam, IAEA, Vienna, (1996).

[25] Cohen L., Schultheiss T.E., Kennaugh R.C. A radiation overdose incident: initial data. Int. J. Radiation Oncology Biol. Phys. Vol 33 (1), (1995) 217-224.

[26] Ash D., Bates T. Report on the clinical effect of inadvertent underdosage in 1045 patients. Clin. Oncol., 6 (4), (1994) 214-226.

[27] Newman H.F. The malfunction "54" accelerator accidents 1985, 1986, 1987, 1990, The medical basis for radiation preparedness II. Clinical experience and follow-up since 1979. Ricks R.C. and Fry S.A. Eds. (Elsevier, New York, 1990) 165-171.

[28] US NRC, US Nuclear Regulatory Commission. Loss of an iridium-192 source and therapy misadministration at Indiana regional Cancer Center, Indiana, Pennsylvania, on November 16, 1992, NUREG-1480, (1993).

[29] Wu De-Chang. Personal communication (1993).

[30] Nénot J.C. Accident grave d'irradiation par une source de césium 137 en Estonie. Revue Générale Nucléaire, n°6, Nov-Déc. 1995, (1995) 476-477.

[31] International Atomic Energy Agency. The radiological accident in Tammiku, Estonia, IAEA, Vienna, (1998).

[32] Gourmelon P., Multon E., Cassagnou H., et al. Preliminary report on the external irradiation accident at Lilo (Georgia) training center. Report IPSN/DPHD/97-05, (1997).

[33] Secretaria de Energia, Minas e Industria Paraestatal. Accidente por contaminacion con cobalt-60, Mexico 1984, CNSNS-IT-001, (1984).

[34] International Atomic Energy Agency. The radiological accident in Goiânia, IAEA, Vienna, (1988, 1998).

[35] International Atomic Energy Agency. Accidental overexposure of radiotherapy patients in San Jose, Costa Rica. IAEA, Vienna, (1998).

[36] International Atomic Energy Agency. Follow-up mission on the accidental overexposure to radiotherapy patients in San José, Costa Rica, IAEA, sous presse.

[37] Nénot J.C., L'Oural nucléaire. Bilan de 50 ans d'une gestion désastreuse. La revue du praticien, Paris, 44, 12, (1994) 1649-1654.

[38] Mettler F.A. and Allen S.N. Military radiation accidents, Medical management of radiation accidents. Mettler F.A., Kelsy C.A. and Ricks R.C. Eds. (CRC Press. 1990) 45-88.