A 60-year-old man found himself in a hospital ward earlier this month after an experiment meant to improve his health instead triggered a rare and dangerous condition known as bromism.
Concerned about the adverse effects of sodium chloride, commonly known as table salt, the man consulted ChatGPT for a healthier alternative. The AI chatbot suggested sodium bromide, a chemical more commonly used in industrial cleaning and pool maintenance, not for human consumption.
Unaware of the danger, he replaced all table salt in his diet with sodium bromide purchased online. After three months on this regimen, he began experiencing extreme thirst, coordination problems, skin changes, and severe psychiatric symptoms, including paranoia and hallucinations.
A report of the man's case was published last week (Aug 5) in the journal Annals of Internal Medicine Clinical Cases.
Background
Bromide toxicity, or bromism, was a common condition in the early 20th century, responsible for up to 8% of psychiatric hospital admissions due to its presence in over-the-counter remedies for insomnia, hysteria, and anxiety. Its incidence fell sharply after the FDA phased out bromide between 1975 and 1989.
In recent years, however, rare cases have resurfaced, linked to dietary supplements, bromide-based sedatives, and excessive dextromethorphan use. With bromide-containing products now more accessible online, experts stress the importance of recognizing bromism as a reversible cause of new psychiatric, neurological, and skin symptoms.
A Rare and Dangerous Diagnosis
A 60-year-old man with no past psychiatric or medical history arrived at the emergency department claiming he believed his neighbor was trying to poison him.
When he was admitted to the hospital, doctors quickly diagnosed bromism, a toxic syndrome caused by excessive bromide exposure. Lab tests revealed pseudohyperchloremia, a false elevation in chloride levels that occurs when bromide interferes with certain tests.
As his hallucinations worsened and he attempted to leave the hospital, medical staff placed him under an involuntary psychiatric hold. Treatment involved fluids, correcting electrolyte imbalances, administering antipsychotic medication, and psychiatric care. Over the course of three weeks, his condition gradually improved.
During his recovery, he also reported additional symptoms such as acne and cherry angiomas, skin changes consistent with bromide toxicity. He was eventually discharged and remained stable during follow-up visits.
Why Blindly Following AI Can Be Dangerous
The incident has sparked concern among medical professionals about the risks of relying on AI chatbots for health advice without verification. Large language models, such as ChatGPT, can generate text that appears authoritative but may be incomplete, decontextualized, or incorrect.
This case highlights a growing trend: an increasing number of people are seeking quick health solutions from AI rather than consulting healthcare providers. Experts warn that such reliance can lead to dangerous outcomes, particularly when dealing with chemicals, supplements, or medications. Even seemingly harmless advice can be risky if applied incorrectly, taken in excessive amounts, or used without understanding possible side effects.
According to OpenAI's service statement, "Our services are not intended for use in the diagnosis or treatment of any health condition. You are responsible for complying with applicable laws for any use of our Services in a medical or healthcare context."
While AI can be a helpful research tool, it should never be treated as a sole authority in matters that affect health and safety.
Leave a Reply