Associated Incidents
AIID editor's note: This peer-reviewed journal article is abridged in parts. See the original source for the complete version, specifically Table 1 and the References section.
Abstract
Ingestion of bromide can lead to a toxidrome known as bromism. While this condition is less common than it was in the early 20th century, it remains important to describe the associated symptoms and risks, because bromide-containing substances have become more readily available on the internet. We present an interesting case of a patient who developed bromism after consulting the artificial intelligence--based conversational large language model, ChatGPT, for health information.
Background
Bromide toxicity, or bromism, was once a well-recognized toxidrome in the early 20th century that precipitated a range of presentations involving neuropsychiatric and dermatologic symptoms (1, 2). Bromism was thought to have contributed to up to 8% of psychiatric admissions at that time, as bromide salts were found in many over-the-counter medications targeting a wide array of indications, including insomnia, hysteria, and anxiety (1, 2). The incidence of bromism declined dramatically when the U.S. Food and Drug Administration eliminated the use of bromide between 1975 and 1989 (3). However, case reports of bromism have reemerged within recent years, including from dietary supplements, bromide-containing sedatives, and excess dextromethorphan (2, 4, 5). While cases of bromism may remain relatively rare, it remains prudent to highlight bromism as a reversible cause of new-onset psychiatric, neurologic, and dermatologic symptoms, as bromide-containing substances have become more readily available with widespread use of the internet.
Case Report
A 60-year-old man with no past psychiatric or medical history presented to the emergency department expressing concern that his neighbor was poisoning him. He initially did not report taking any medications, including supplements. His vital signs and physical examination, including neurologic examination, were normal. The initial laboratory evaluation was notable for hyperchloremia (126 mmol/L; normal range, 98 to 108), a negative anion gap (--21 mEq/L), and a low phosphate level (<1 mg/dL; normal range, 2.5 to 4.5). His bicarbonate level was elevated (36 mEq/L) and his venous blood gas test revealed a compensated respiratory acidosis and metabolic alkalosis (pH, 7.35; Pco2, 64 mm Hg; Po2, 23 mm Hg; bicarbonate, 35 mEq/L). Sodium (141 mEq), creatinine (0.97 mg/dL), urine drug screen, and blood alcohol levels were normal (Table 1). He was admitted to a medical telemetry bed for electrolyte monitoring and repletion.
On admission, the patient shared that he maintained multiple dietary restrictions and that he distilled his own water at home. He was noted to be very thirsty but paranoid about water he was offered. A broad differential, including heavy metal ingestion, was considered, prompting consultation with the Poison Control Department. Based on the differential for a negative anion gap with a normal sodium level, this was thought to be a case of pseudohyperchloremia (6). A normal salicylate level and fasting lipids left bromism as the most likely cause per discussion with Poison Control colleagues and review of UpToDate (6). In the first 24 hours of admission, he expressed increasing paranoia and auditory and visual hallucinations, which, after attempting to escape, resulted in an involuntary psychiatric hold for grave disability. He received risperidone, which was titrated up to 3 mg daily for psychosis.
After treatment with intravenous fluids and electrolyte repletion, he became medically stable for admission to the inpatient psychiatry unit. His metabolic alkalosis resolved and was thought to be due to chloride depletion and compensatory for his respiratory acidosis of unclear cause. His hypophosphatemia was thought to be from refeeding syndrome as the patient described an extremely restrictive vegetarian diet and was found to have multiple micronutrient deficiencies, including vitamin C, B12, and folate deficiencies. Vitamin D levels were not tested. With improvement, he was able to report that he had recently noticed new-onset facial acne and cherry angiomas, fatigue, insomnia, subtle ataxia, and polydipsia, further suggesting bromism. He also shared that, after reading about the negative effects that sodium chloride, or table salt, has on one's health, he was surprised that he could only find literature related to reducing sodium from one's diet. Inspired by his history of studying nutrition in college, he decided to conduct a personal experiment to eliminate chloride from his diet. For 3 months, he had replaced sodium chloride with sodium bromide obtained from the internet after consultation with ChatGPT, in which he had read that chloride can be swapped with bromide, though likely for other purposes, such as cleaning.
Gradually, over the course of a 3-week admission, his chloride and anion gap normalized and psychotic symptoms improved. He was tapered off risperidone before discharge and remained stable off medication at a check-in 2 weeks after discharge. His bromide level ultimately was 1700 mg/L (21 mmol/L; reference range, 0.9 to 7.3 mg/L).
Discussion
Despite the limited use of bromide in ingestible products today, cases of bromism persist, as awareness of the risk fades and availability of medications and supplements containing bromide increases on the internet. Thus, it remains important to consider bromism in patients presenting with new neurologic, psychiatric, and/or dermatologic symptoms, as well as hyperchloremia with a negative anion gap (2). Our institution's laboratory uses the commonly utilized ion selective electrode (ISE) assay to measure serum chloride. The ion selective electrode assay is known to give falsely elevated chloride readings in the setting of elevated levels of other halides such as bromide but not at lower physiologic levels (1). The dedicated bromide level was determined via inductive coupled plasma mass spectrometry. Diagnosis is critical, as toxicity may be completely reversible with cessation of bromide ingestion and treatment, involving aggressive saline diuresis with intravenous normal saline (7).
This case also highlights how the use of artificial intelligence (AI) can potentially contribute to the development of preventable adverse health outcomes. Based on the timeline of this case, it appears that the patient either consulted ChatGPT 3.5 or 4.0 when considering how he might remove chloride from this diet. Unfortunately, we do not have access to his ChatGPT conversation log and we will never be able to know with certainty what exactly the output he received was, since individual responses are unique and build from previous inputs.
However, when we asked ChatGPT 3.5 what chloride can be replaced with, we also produced a response that included bromide. Though the reply stated that context matters, it did not provide a specific health warning, nor did it inquire about why we wanted to know, as we presume a medical professional would do.
Thus, it is important to consider that ChatGPT and other AI systems can generate scientific inaccuracies, lack the ability to critically discuss results, and ultimately fuel the spread of misinformation (8, 9). While it is a tool with much potential to provide a bridge between scientists and the nonacademic population, AI also carries the risk for promulgating decontextualized information, as it is highly unlikely that a medical expert would have mentioned sodium bromide when faced with a patient looking for a viable substitute for sodium chloride (10). Thus, as the use of AI tools increases, providers will need to consider this when screening for where their patients are consuming health information.