Accidents with radioactive sources in industry and hospitals

A 38-year-old man who makes a living collecting scrap metal arrives at the emergency department with nausea, vomiting and diarrhea since the day before. He reports that three days earlier he found a very heavy metal cylinder in an empty lot, took it home and tried to open it with a saw and a blowtorch to sell the lead. Inside there was a small, shiny piece, which he kept in his pants pocket to show the owner of the scrapyard. Now the skin on his right thigh is red and burning, and the palm of his right hand has a swollen and painful area. His wife and 10-year-old son also vomited. The doctor on duty first thinks of food poisoning. The scene is fictional and was written for teaching.

This text uses the case as a clinical reasoning scenario for students and health professionals. The topic is accidents with radioactive sources that get out of control, in industry and in hospitals, and end up in the hands of people who do not know what they are. The focus is what the health team decides: suspect radiation when symptoms are common, notify the authority, protect other people, estimate the dose and treat the skin and bone marrow. It does not replace your service's radiological emergency plan or the guidance of the regulatory authority. If you play the case of the day on NEXO, you will recognize the challenge: deciding with clues that arrive in sequence.

Key points

• Radiotherapy and industrial gamma radiography sources that are abandoned or stolen have already caused deaths among the public in several countries.
• Nausea, vomiting and diarrhea without a clear cause, together with a burn without heat, should raise suspicion of radiation.
• In the accidents analyzed by the IAEA, the first doctors almost never recognized the cause.
• The doctor who suspects must notify the regulatory authority, because the source may still be exposing other people.
• The drop in lymphocytes and chromosome analysis help estimate the dose, but management is guided by clinical signs.
• Radiation skin injuries appear late and worsen for weeks, and dosimetry-guided surgery helps in severe cases.

Why this is a reasoning scenario

The complaint in the scene is one of the most common in an emergency department. Vomiting and diarrhea point to gastroenteritis, and the red skin looks like a blowtorch burn. Clinical reasoning lies in putting together three clues that do not match these hypotheses: a heavy metal object of unknown origin, a skin injury at the spot where the object was kept and several sick people in the same house. This combination has appeared in almost all accidents with lost sources described by the IAEA.

Dörr and Meineke recall that the initial symptoms of acute radiation syndrome include anorexia, nausea, vomiting, diarrhea, fluid loss, fever, hypotension, headache and early erythema (Dörr and Meineke, 2011). They warn that these symptoms may be interpreted as nonspecific gastrointestinal or infectious disease if the possibility of radiation exposure is not considered (Dörr and Meineke, 2011). Iddins and colleagues note that affected people usually belong to three groups: workers exposed in industrial accidents, patients in medical practice accidents and the general public in accidents with uncontrolled sources, also called orphan sources (Iddins and colleagues, 2022).

The post on acute radiation syndrome and lymphocyte triage details the lymphocyte count, and the post on radiation skin injury details the skin progression. This text brings the two together at the moment when no one yet knows that an accident has occurred.

What the scene teaches

The scene trains six habits. The first is to ask about metal objects found or bought as scrap when there are gastrointestinal symptoms without a clear cause. The second is to recognize a burn without a history of heat. The third is to notify the authority before finishing the diagnosis, because the source is still emitting. The fourth is to draw blood early and repeat the count. The fifth is to treat the skin with the idea that the injury will worsen. The sixth is to look for other exposed people, in the family, at the scrapyard and in the neighborhood.

What is not at stake

The text does not cover internal contamination by cesium or Prussian blue, which are in the post about internal contamination by cesium 137. It also does not cover accidents at nuclear power plants or the protection of the interventional radiology team. The focus is the sealed source that left its place and the care of the people who had contact with it.

How a source gets out of control

High-activity radioactive sources have legitimate, controlled uses. Radiotherapy uses cobalt 60 to treat cancer, and industry uses iridium 192 and cobalt 60 to radiograph welds and parts. The problem starts when the source leaves the control system, through abandonment, a change of address without notice, unsafe storage or theft. IAEA reports show the same pattern in different countries.

Goiânia, 1987

At the end of 1985, a private radiotherapy institute in Goiânia moved to a new address, took a cobalt 60 unit with it and left a cesium 137 teletherapy unit in the old building, without notifying the authority, as the license required (IAEA, Goiânia). The building was partly demolished, and two people went in, not knowing what the machine was, and removed the source thinking it had value as scrap (IAEA, Goiânia). While trying to take the piece apart, they broke the capsule, and the highly soluble cesium chloride spread (IAEA, Goiânia). The remains were sold to a scrapyard, and the owner noticed that the material glowed blue in the dark (IAEA, Goiânia). Fragments the size of rice grains were given to several families (IAEA, Goiânia).

The symptoms were not recognized at first as an effect of radiation (IAEA, Goiânia). One of the affected people linked the illnesses to the capsule and took the remains to the city public health service, which made it possible to identify the accident (IAEA, Goiânia). In total, about 112,000 people were monitored, and 249 had internal or external contamination (IAEA, Goiânia). Four people died in the first four weeks of hospitalization, with hemorrhagic and septic complications of acute radiation syndrome (IAEA, Goiânia).

Tammiku, 1994

On October 21, 1994, three brothers entered a radioactive waste depot in Tammiku, Estonia, without authorization and removed a metal container with a cesium 137 source (IAEA, Tammiku). The source fell to the floor, and one of the men put it in his pocket and took it home (IAEA, Tammiku). He began to feel ill soon after, vomited a few hours later and died on November 2, with severe injuries to his leg and hip (IAEA, Tammiku). The death was not attributed to radiation, and the source stayed in the house with the man's wife, stepson and the boy's great-grandmother (IAEA, Tammiku). On November 17, the boy was admitted with severe burns on his hands, and a doctor identified them as radiation injuries (IAEA, Tammiku).

Lilo, 1997

In October 1997, the government of Georgia informed the IAEA that nine soldiers from a training center in Lilo had radiation-induced skin injuries on various parts of the body (IAEA, Lilo). The search found 12 cesium 137 sources, one cobalt 60 source and 200 radium 226 sources (IAEA, Lilo). One of the sources was taken from the pocket of a soldier's winter coat on September 13, 1997 (IAEA, Lilo). Later, two other people were identified as exposed, and all 11 had cutaneous radiation syndrome (IAEA, Lilo).

Istanbul, 1998, and Samut Prakarn, 2000

In Istanbul, a licensed company kept used radiotherapy sources for years in transport packages that should have gone back to the supplier (IAEA, Istanbul). The packages ended up in a property that changed owners, and the new owners sold them as scrap (IAEA, Istanbul). The buyers opened the shielded containers on December 10, 1998, in a residential area, and exposed themselves to at least one unshielded cobalt 60 source (IAEA, Istanbul). In Samut Prakarn, Thailand, a company without authorization kept teletherapy heads in a place with no security (IAEA, Samut Prakarn). On February 1, 2000, two people took a partially dismantled head to a scrapyard, and the source fell out of the casing while a worker cut the metal with a blowtorch (IAEA, Samut Prakarn). The head had the radiation symbol and a warning label, but the people did not know what the symbol meant, and the label was not in a language they understood (IAEA, Samut Prakarn).

The hospital as the source of the problem

The hospital appears in these accounts in two ways. The first is as the origin of the source, because the teletherapy equipment in Goiânia, Istanbul and Samut Prakarn came from radiotherapy (IAEA, Goiânia) (IAEA, Istanbul) (IAEA, Samut Prakarn). The second is as the place of accidental exposure of the patient. In Białystok, Poland, on February 27, 2001, a power failure shut down a linear accelerator during treatment, and the machine was switched back on (IAEA, Białystok). Five patients were treated next, and two felt itching and burning during irradiation, which led the team to stop (IAEA, Białystok). Measurements showed that the machine output was far above what was expected, because the dose monitoring system was not working and a component of the safety interlock was damaged (IAEA, Białystok). The five patients had local skin reactions of varying severity (IAEA, Białystok).

Signs that raise suspicion

Suspicion comes from the combination of signs, not from a single sign. Each of them has more common explanations, and it is the sum that points to radiation. The doctor at the scene has the pieces in hand, but needs to ask about the origin of the cylinder.

The general signs

In Samut Prakarn, the exposed people went to hospitals and outpatient clinics, but the first symptoms, diarrhea, nausea and vomiting, were not initially attributed to radiation (IAEA, Samut Prakarn). Many of them went to an outpatient clinic with weakness, hair loss and burns, along with a history of nausea, vomiting or diarrhea (IAEA, Samut Prakarn). In Istanbul, ten people who spent time near the dismantled containers fell ill on December 13, 1998, and six began to vomit (IAEA, Istanbul). They sought care, but the cause was only recognized almost four weeks later, on January 8, 1999 (IAEA, Istanbul).

The signs that, together, should lead to the question about radiation are these:

• Nausea, vomiting or diarrhea in several people from the same household or the same workplace, with no common suspect food.
• Redness, blister or wound on the hand, thigh or hip with no history of heat, chemical product or friction.
• Skin lesion at the point where an object was kept, such as a trouser or coat pocket.
• Hair loss, bleeding gums or nosebleeds and fever in the following weeks.
• Recent contact with scrap, parts of medical or industrial equipment or very heavy metal containers.
• Mention of a shiny part, a trefoil symbol or warning labels.

The skin tells the story

Iddins and colleagues describe that the patient may have transient erythema a few hours after irradiation, linked to changes in vessel permeability, and that the time when this erythema appears has prognostic value (Iddins and colleagues, 2022). Lower doses cause transient lesions, and doses above 25 Gy lead to tissue necrosis in the following weeks or months (Iddins and colleagues, 2022). In accidents with a skin dose above 25 Gy, the sequence was transient erythema, late erythema, dry and moist desquamation, ulcer and necrosis over a period of one to five weeks (Iddins and colleagues, 2022).

The table by Iddins and colleagues gives the absorbed dose thresholds in the skin for each manifestation (Iddins and colleagues, 2022).

Skin manifestations	Absorbed dose threshold
Erythema of the second phase	3 Gy
Temporary hair loss	3 Gy
Permanent hair loss	7 Gy
Dry desquamation	10 Gy
Wet desquamation	15 Gy
Necrosis	25 Gy or more

The skin does not tell the whole story. In Lilo, exposure was very heterogeneous and spread over a long time, and the patients had skin reactions at different stages between April and August 1997 (IAEA, Lilo). The report notes that the likely exposure time to the radiation field was between 60 and 300 days (IAEA, Lilo). Even with lesions only on the skin, the cesium 137 source, which emits mostly gamma rays, also exposed internal organs, such as the bone marrow and the reproductive system (IAEA, Lilo). Signs of whole-body exposure, such as absence of sperm in the semen or altered follicle-stimulating hormone, appeared in 8 of the 11 patients (IAEA, Lilo).

First steps in care

The first step after suspicion is to protect people. The source may be in the patient's pocket, in his home or at the scrapyard, and every hour counts for those who remain near it. The second step is to stabilize the patient, because no sealed source makes the patient dangerous to the team if the source is not with him.

Notify the authority

In Istanbul, when the lesions were finally attributed to radiation, the doctor immediately notified the national authorities, and an unshielded source was soon found at the scrapyard and recovered safely, which prevented further exposure (IAEA, Istanbul). The report recommends that the doctor contact the regulatory authority when suspecting that the patient's symptoms may be radiation-induced (IAEA, Istanbul). In Samut Prakarn, doctors at a local hospital recognized the signs and symptoms and suspected an uncontrolled source, and notified the regulatory authority (IAEA, Samut Prakarn). The response team found high radiation levels near the scrapyard, isolated the area and recovered the source by February 20, 2000 (IAEA, Samut Prakarn). The activity of the recovered source was estimated at 15.7 TBq of cobalt 60 (IAEA, Samut Prakarn).

In Tammiku, only the attention of the doctor who examined the boy on November 17, 1994 made it possible to recognize the cause of the lesion (IAEA, Tammiku). The report states that this medical reaction was decisive, because it started the rescue, limited the consequences for the family and prevented accumulated doses in the neighboring population (IAEA, Tammiku). The source was recovered from the house and returned to the depot on November 18 (IAEA, Tammiku).

What to do in the first hour

The measures for the first hour, according to the reports, follow this order:

• Ask whether the object or part of it is with the patient, in his clothing or in his bag, and do not touch suspected metal pieces with your hand.
• Notify the radiation regulatory authority and health surveillance as soon as suspicion arises.
• Record the likely time contact began and the time vomiting began.
• Collect a blood count with lymphocyte count as soon as possible and repeat at intervals.
• Set aside a few milliliters of blood for biological dosimetry before any transfusion.
• Photograph and mark the skin lesions, with date and time.
• List the people who had contact with the object, at home, in transport and at the scrapyard.

The Istanbul report recommends collecting blood for hematological and cytogenetic evaluation as soon as possible and repeating the collection at intervals, to estimate the dose and the prognosis (IAEA, Istanbul). It also recommends saving a few milliliters of blood for biological dosimetry before whole blood transfusions (IAEA, Istanbul). The health team does not need to be expert in physics to do this, but it needs to know that this sample is lost afterward.

Protection of the response team

Protecting those who recover the source is the responsibility of the specialized team. In Samut Prakarn, the people who worked in the response and recovery received whole-body doses that did not exceed 32 mSv during the entire operation (IAEA, Samut Prakarn). The emergency department health team should not try to collect the source, but rather keep people away from the object and wait for the authority's team.

Dose estimation and treatment

The dose defines the prognosis and the intensity of treatment, but it is almost never known at the start. The reasoning combines the clinical picture, the blood count, cytogenetics and the physical reconstruction of the accident.

Biological and physical dosimetry

Dörr and Meineke explain that, since physical dosimetry and biodosimetry are almost never available in an accident, the estimate of effects and the prognosis are based on clinical signs and symptoms, as in the METREPOL system (Dörr and Meineke, 2011). This system grades the effects in four systems: neurovascular, hematopoietic, cutaneous and gastrointestinal (Dörr and Meineke, 2011). Lima and colleagues report that the count of dicentric chromosomes in lymphocytes has a linear quadratic dose response relationship up to 5 Gy in acute photon exposure and detects whole body doses from about 0.1 Gy (Lima and colleagues, 2015). They describe that the thresholds for irradiation within up to two days are more than 1 Gy for the hematopoietic form, 6 to 10 Gy for the gastrointestinal form and more than 20 Gy for the cerebrovascular form (Lima and colleagues, 2015).

Prolonged exposure complicates the calculation. In Tammiku, the whole body dose estimate from the acute dicentric calibration curve, of 1.1 Gy, clearly underestimated the dose, and the estimate for prolonged exposure, of 2.7 Gy, was more consistent with the blood count (IAEA, Tammiku). The report concludes that, when prolonged exposure is expected, as in the case of a lost source, the team must be prepared for a more severe and longer drop in blood cells than in an acute exposure of similar estimated dose (IAEA, Tammiku). It also recommends planning treatment by symptoms, and not only by the estimated dose (IAEA, Tammiku).

In Istanbul, the samples for biodosimetry were collected one month after the accident, and many lymphocytes with dicentrics had already been cleared from the blood (IAEA, Istanbul). The FISH translocation method proved useful for estimating the dose when collection was late (IAEA, Istanbul). In Samut Prakarn, there was no adequate biological dosimetry, and the report recommends that each country with sources capable of causing serious accidents designate a cytogenetics laboratory with its own curves for cobalt 60, iridium 192 and cesium 137 (IAEA, Samut Prakarn).

For local injury, physical reconstruction helps decide on surgery. Lima and colleagues describe the use of electron paramagnetic resonance in tooth enamel, nails and other materials, and of Monte Carlo simulation with body models, to build dose maps (Lima and colleagues, 2015). In the 2005 accident in Chile, the simulation done in France estimated a maximum local dose of 2,000 Gy at the center of the lesion, with a very rapid fall in dose with depth and at the surface (Lima and colleagues, 2015).

Marrow treatment and infection prevention

Dörr and Meineke describe as principles of treatment of the hematopoietic form the replacement of blood products, the use of cytokines such as G-CSF and GM-CSF and hematopoietic stem cell transplantation (Dörr and Meineke, 2011). They recommend starting G-CSF or GM-CSF early when severe neutropenia is expected (Dörr and Meineke, 2011). In Samut Prakarn, with adequate support, platelet transfusions and stimulation with G-CSF and GM-CSF, many victims reached safe counts, and marrow transplantation was not needed, even with severe hematopoietic syndrome from a dose of several grays (IAEA, Samut Prakarn).

Even so, three people died of infectious complications within two months (IAEA, Samut Prakarn). The report recommends considering immediate reverse isolation and antivirals, antibacterials and antifungals when severe whole body exposure is suspected (IAEA, Samut Prakarn). In Istanbul, strict reverse isolation of the ten people admitted with acute syndrome and the broad use of antimicrobials were considered justified to prevent hospital infection (IAEA, Istanbul). G-CSF, started on January 12, 1999, led to a rapid recovery of leukocytes in 4 to 11 days (IAEA, Istanbul).

A very high dose changes the picture. Guo and colleagues describe a 32-year-old man exposed to 14.5 Gy of gamma rays in a cobalt 60 accident in China in 2008 (Guo and colleagues, 2014). He received a stem cell and mesenchymal cell transplant and had hematopoietic recovery, but died on day 62 of multiple organ failure linked to gastrointestinal damage (Guo and colleagues, 2014). The authors note that the literature describes death within 60 days for all patients classified as H4 by the METREPOL system (Guo and colleagues, 2014).

Skin treatment

Iddins and colleagues state that severe cases of local injury, with radionecrosis, will probably require admission to plastic and reconstructive surgery or burns services with experience in this type of injury (Iddins and colleagues, 2022). They describe dosimetry-guided surgery, applied early in cases of severe local exposure, in which the dose map shows the areas above 25 Gy where necrosis is likely (Iddins and colleagues, 2022). Mesenchymal cells can be injected during surgery and in subsequent sessions (Iddins and colleagues, 2022). Ultrasound and thermography help to see changes beneath the skin that do not yet appear on clinical examination (Iddins and colleagues, 2022).

In the Chile accident, in 2005, a worker found an iridium 192 source from industrial radiography on a scaffold and kept it in his back trouser pocket while he took it to the supervisor (Iddins and colleagues, 2022). He had severe injuries to his hands and buttock, treated with dosimetry-guided surgery and mesenchymal cells (Iddins and colleagues, 2022). The scene in this text repeats that pattern: the hand that held it and the thigh that stayed next to the pocket.

Risks and reasoning errors

The reports repeat the same mistakes, and almost all of them are errors of reasoning, not of technique. Knowing how to recognize them helps avoid the delay that made some accidents more severe.

The most frequent mistakes are closing the diagnosis of gastroenteritis without asking about objects or scrap, calling the injury a thermal burn without a history of heat, waiting for confirmation to notify the authority while the source continues to expose the family, relying on a single dose estimate in prolonged exposure, collecting blood for cytogenetics too late, and operating on the skin without knowing how far the dose extends.

Anchoring on the common diagnosis

The Istanbul report notes that, in many accidents, doctors fail to recognize a radiation injury, and that early diagnosis and treatment can save lives (IAEA, Istanbul). In Lilo, treatment was more difficult because of the late diagnosis (IAEA, Lilo). In Tammiku, the report recalls that accidents with lost or stolen sources can easily be misinterpreted, and that the cause was often discovered by chance (IAEA, Tammiku). The post about diagnostic pause helps to stop before closing the picture as food poisoning.

One family, several patients

The scene has more than one patient. In Samut Prakarn, those who picked up the source had severe local injuries, but whole-body doses of about 2 Gy, while those who stayed at the scrap yard received higher doses from prolonged exposure (IAEA, Samut Prakarn). Four of these people received more than 6 Gy, and three of them died in the following two months (IAEA, Samut Prakarn). The person who looks sickest on arrival is not necessarily the one who received the highest dose. The post about interprofessional clinical reasoning shows why the physicist, the hematologist and the plastic surgeon need to get involved early.

Fear and mass seeking

An accident like this also floods services with worried people. In Istanbul, 404 people asked for medical or hematological tests for fear of having been exposed, and 18, among them seven children, were admitted (IAEA, Istanbul). Ten adults had clinical signs of acute radiation syndrome, and the five most severe stayed in hospital for 45 days (IAEA, Istanbul). Triage must separate those who had real contact from those who are only afraid, without dismissing either group. The post about diagnostic uncertainty helps you decide with incomplete data.

How NEXO helps train this reasoning

NEXO publishes one fictional clinical case per day, with up to six progressive clues. You can attempt the diagnosis at each clue, searching for the code in ICD-10, and you get proximity feedback at five levels. The format trains the same skill as this scene: putting together data that arrives little by little, such as the vomiting, the lesion on the pocket, the metal cylinder and the sick people in the same house, and deciding without closing too early. The post about case-based learning explains why this format works, and the post about clinical decision-making helps you choose the order of the steps.

Three exercises with cases

The first exercise is to replay the scene with an industrial radiography technician who arrives with pain and a blister on the fingers after a work shift and decide which questions to ask. The second is to replay the scene with a radiotherapy patient who feels burning during the session and decide what the team should do with the machine. The third is to replay the scene with a child who has a burn on the hand and whose father died days earlier with lesions on the leg.

Teachers and content creators can build their own cases in the space for Creators. The exercise works well in internal medicine, emergency medicine, hematology and occupational medicine residencies and in public health surveillance training.

Frequently asked questions

Is a person exposed to a sealed source dangerous to the team?

No, if the source is not with them. The risk comes from the source, and that is why the first question is where the object is (IAEA, Istanbul).

Who should be notified when suspicion arises?

The radiation regulatory authority, to ensure public safety (IAEA, Istanbul).

When does skin necrosis appear?

With doses above 25 Gy, in the weeks or months after exposure (Iddins et al., 2022).

Is bone marrow transplantation always indicated?

No. In Samut Prakarn and in Istanbul, supportive care and G-CSF were enough for doses of a few grays (IAEA, Samut Prakarn) (IAEA, Istanbul).

Why can the estimated dose be misleading?

Because in prolonged exposure the acute curve underestimates the dose, and treatment should follow the symptoms (IAEA, Tammiku).

Final checklist

• I asked about metal objects, scrap or equipment found.
• I looked for a skin lesion without a history of heat and at the pocket site.
• I confirmed that the source is not with the patient.
• I notified the regulatory authority and public health surveillance.
• I noted the time of contact and the time of the first vomiting.
• I ordered a blood count with lymphocytes and saved blood for cytogenetics before transfusions.
• I considered reverse isolation and G-CSF in the face of severe exposure.
• I referred the severe skin lesions to a service with experience.
• I listed the family, the scrapyard coworkers and the exposed neighbors.

Want to practice reasoning with progressive clues? Play today's case.

References

1. Dörr H, Meineke V. Acute radiation syndrome caused by accidental radiation exposure - therapeutic principles. BMC Med. 2011. PMID 22114866.
2. Guo M, Dong Z, Qiao J, et al. Severe acute radiation syndrome: treatment of a lethally 60Co-source irradiated accident victim in China with HLA-mismatched peripheral blood stem cell transplantation and mesenchymal stem cells. J Radiat Res. 2014. PMID 23979075.
3. Iddins CJ, DiCarlo AL, Ervin MD, et al. Cutaneous and local radiation injuries. J Radiol Prot. 2022. PMID 34488201.
4. International Atomic Energy Agency. The Radiological Accident in Goiania. Vienna: IAEA, 1988. Official source.
5. International Atomic Energy Agency. The Radiological Accident in Tammiku. Vienna: IAEA, 1998. Official source.
6. International Atomic Energy Agency, World Health Organization. The Radiological Accident in Lilo. Vienna: IAEA, 2000. Official source.
7. International Atomic Energy Agency. The Radiological Accident in Istanbul. Vienna: IAEA, 2000. Official source.
8. International Atomic Energy Agency. The Radiological Accident in Samut Prakarn. Vienna: IAEA, 2002. Official source.
9. International Atomic Energy Agency. Accidental Overexposure of Radiotherapy Patients in Białystok. Vienna: IAEA, 2004. Official source.
10. Lima CM, Lima AR, Degenhardt ÄL, et al. Reconstructive dosimetry for cutaneous radiation syndrome. Braz J Med Biol Res. 2015. PMID 26445332.