Reconsidering the Seveso Dioxin Disaster

by Brian Dunning
January 30, 2024

Today we're going to point the skeptical eye at a 1976 industrial disaster at a chemical plant in Seveso, Italy, long considered one of the worst in history. Thousands of local residents were exposed to dioxin, widely believed to be one of the most toxic chemicals there is. We're going to see how bad it really was, and whether those labels are really deserved. Developing effective solutions depends on having an accurate understanding of a problem; so if the way we regard Seveso today is inaccurate, then we're setting ourselves up to fail when we respond to some future similar event.

So what is dioxin, anyway? Surprisingly this is not a simple question. There are hundreds of chemicals that are classified as dioxins, falling into three groups: PCDDs, PCDFs, and PCBs. All dioxins are persistent organic pollutants, meaning they take a long time to break down in the environment. But only some of these are toxic, those that have a chlorine atom in certain positions in the molecule. Those that are are considered highly toxic. They are produced as unwanted byproducts during some chemical production processes, including the manufacturing of some herbicides. But they are also produced naturally by combustion such as forest fires and volcanic eruptions; thus, they are (and always will be) found throughout the environment. About 90% of human exposure is via the consumption of animal fats, basically meat and dairy. Dioxin stays in the fat of the animals and of the people who eat them, where its half life can be up to a decade. Since much of this dioxin in the environment comes from natural sources it's impossible for us to keep it out, but industry and regulators have been working together to eliminate the human-caused sources for a long time.

The dioxin at Seveso was a PCDD called TCDD (I'll spare you the complete chemical name because I don't feel like pronouncing it and it doesn't matter for our purposes today). TCDD is what's usually meant when we colloquially use the word dioxin — going forward in this episode, when I say dioxin, I'm referring to TCDD. Dioxin is not good. It's listed as a carcinogen and a cause of birth defects and about everything else; however, the evidence is strong only for two health effects: chloracne (a type of environmentally triggered skin lesion resembling acne, which can be severe) and transient mild hepatotoxicity (temporary mild impairment to liver function). It might (and probably does) do worse, but these are the only confirmed effects. The fact is we just don't know. But regardless, you want as little dioxin in your body as possible beyond what the natural environment is going to give you as a baseline.

Dioxin is also the component of Agent Orange believed to be responsible for negative health impacts associated with its use in Vietnam. Agent Orange probably warrants its own episode for another time, but for now, all we're going to say is that dioxin has these two known effects, and possibly others. Suspected effects include cancer, reproductive and developmental problems, damage to the immune system, and endocrine disruption.

So on July 10, 1976, a chemical manufacturing plant in Italy owned by a company called ICMESA had an incident. A string of human errors led a chemical compound to increase in pressure and temperature over the weekend when operations were supposed to be closed. A relief valve blew, releasing six tons of chemicals into the atmosphere, which settled over an area estimated at 18 square kilometers. This cloud was mostly composed of sodium hydroxide, ethylene glycol, and sodium trichlorophenate — themselves not much of a danger in that form — but also probably contained between 1 and 30 kg of dioxin.

As this was one of the most extensively studied industrial disasters in history, we have a lot of pretty detailed information on the effects of that dioxin. Those 18 km² were divided into three zones. Zone A, with about 750 residents, is where soil concentration of dioxin was above 50 µg/m²; Zone B with a bit less than 5,000 people was less than 50 µg/m²; and Zone R, with over 30,000 residents, had a negligible concentration.

Dioxin is known to be more toxic to many animals than it is to humans, and the cloud killed about a quarter of all animals in Zone A immediately. Within a few weeks, 3,300 animals were dead. Eventually, 80,000 animals were culled to prevent them entering the food chain. As far as people go, only those in Zone A were evacuated. Everyone in all three zones was cautioned not to eat any locally sourced food. People in Zones A and B were given medical testing.

Due to Italian officials' initial denials that anything much had happened, it was two weeks before the event was reported in the rest of the world. Among the first was London's Daily Telegraph on July 23, with the headline "Poison gas cloud over town after blast":

Dogs, cats, rabbits and fowl in the area began to die immediately after the blast. Experts said later than just 10 per cent. of the chemical which escaped would be enough to kill the 11 million population of New York if mixed with the water supply.

The New York Times didn't report on it until July 29, with the headline "20 More Evacuated From Area in Italy Hit by Poison":

The most significant substance that escaped from the chemical plant was said by an American expert to be dioxin, one of the deadliest chemicals known. A dose of less than a billionth of a gram is fatal to guinea pigs.

The amount believed to have escaped from the Italian factory... is said to be about four and a half pounds. No one knows what the lethal dose is for human beings but no one doubts that four and a half pounds would kill many thousands of people.

With such shocking reports, no doubt the actual loss of life must have been enormous. Let's find out.

There's a particularly infamous case of dioxin poisoning that you may remember (if you're "of a certain age"). In September 2004, Ukrainian Presidential candidate Viktor Yushchenko was poisoned with a massive dose of TCDD dioxin at a dinner event, during Ukraine's Orange Revolution. He became severely ill and was hospitalized with 50,000 times as much dioxin in his system as normal. Afterward, his only lasting symptom was severe, disfiguring chloracne making him nearly unrecognizable. The chloracne was so severe that his face scarred, making his altered appearance permanent. Yushchenko's case, however, cannot really be compared to Seveso, because he actually ingested a huge amount; while the Seveso victims had skin and inhalation exposure to a far, far lower concentration. Yushchenko's case was so uniquely severe that scientific papers were written about him, and he provided tissue samples for years; and his case was instrumental in guiding our knowledge of dioxin's long term toxicokinetic effects in the body. (Incidentally, he won the election, bringing the Orange Revolution to an end, and was inaugurated just four months after he was poisoned.)

In the end, the final adverse health effects to the Seveso population was 193 cases of chloracne, mostly among children, all but one of whom were considered completely recovered by 1983. Nobody was killed. A number of papers have been published attempting to characterize the total impact; none found sufficient evidence to definitively link anything else to the dioxin exposure. Here's from one such paper:

​​Other reversible early effects noted were peripheral neuropathy and liver enzyme induction. The ascertainment of other, possibly severe sequelae of dioxin exposure (e.g., birth defects) was hampered by inadequate information; however, generally, no increased risks were evident. Mortality studies shed some light on the long-term effects. An unusual cardiovascular mortality pattern was reported in the exposed population. Cancer mortality findings after 10 years do not allow firm conclusions to be drawn, but are suggestive of a departure from expectations for certain types of cancer; the ongoing cancer incidence study will further explore these hypotheses.

Yet, Seveso made #8 on TIME Magazine's 2010 article "And the Earth Cried: Top 10 Environmental Disasters." Was it an environmental disaster as well as a public health disaster? It doesn't really appear so. The total environmental cleanup consisted of bulldozing tons of soil, debris from demolished homes that were considered the most contaminated, and animal remains into two piles covered with concrete, and which now form hills in a public park; in addition, 42 barrels containing waste chemicals from the plant and used protective clothing were taken away for incineration. No other cleanup was deemed necessary.

We should note that TIME's list also includes Three Mile Island, the supposed 1979 nuclear disaster in Pennsylvania. In that, nobody was killed or injured, no significant amount of radioactive material escaped, the safety systems functioned as intended, and the plant resumed normal operations. But rather than being showcased as a success story, Three Mile Island is almost always portrayed as a disaster. Why?

Sometimes we tend to have a strange way of categorizing what's a disaster and what's not. Three Mile Island, which hurt nobody, is usually considered a catastrophe; while the coal power industry, which directly kills hundreds of thousands of people around the world every year from lung cancer, is not. The Texas City Disaster of 1947 killed 581 people, yet it's not on the list either. Perhaps we should recalibrate our scale of "worst disasters."

Quite obviously, when I imply something like "Hey, Seveso wasn't all that bad, all they got was acne; look at this list of things that were worse" it sounds incredibly dismissive, almost like I'm speaking in support of the chemical company that caused it. What I'm speaking against, in fact, is the common tendency to have a disproportionate reaction to events that involve scary words like dioxin or chemical or nuclear when far more harmful events don't have nearly as much cultural influence, simply because they lack the triggering buzzwords. The result is that attention and resources get misallocated. Just think how much better off the world would be if the amount of media and cultural attention given to the Fukushima nuclear meltdown, which produced minimal casualties, was directed instead at the coal power industry, with its ongoing death toll up to perhaps the tens of millions.

We are emotional animals, to be sure. The challenge to all of us is to allow data to have the amount of influence on our behavior that it deserves as well.

By Brian Dunning
Follow @BrianDunning