Bad Advice
Want to know what poisoned advice was given by a chatbot? What drug toxicity is also a euphemism for trite or boring? What toxin is in the Dead Sea?
This is the Pick Your Poison podcast. I’m your host Dr. JP and I’m here to share my passion for poisons in this interactive show. Will our patient survive this podcast? It’s up to you and the choices you make. Our episode today is called Bad Advice. Want to know what poisoned advice was given by a chatbot? What drug toxicity is also a euphemism for trite or boring? What toxin is in the Dead Sea? Listen to find out.
Today's episode starts at your parents’ house. You’re scheduled for the night shift and came over for a hearty dinner beforehand. You’re sitting in the kitchen. Your dad is making your favorite dishes while your mom tells you how much she’s enjoying retirement. The conversation is interrupted by pounding on the front door. It sounds like someone is trying to knock it down, rather than just announce their arrival. Your parents exchange a glance and say just ignore it. It eventually stops and you sit down to enjoy the meal.
Halfway thru, it starts again. You get up to see what is happening. Your mom waves you back saying it's just the neighbor. He's losing his mind. You peer out the window. The neighbor, an elderly man in his 60s is beating on the door with his fists shouting for your parents to stop poisoning his water.
He stops again, only to restart again 30 minutes later. As before, he’s shouting and accusing them of poisoning his water. Your parents’ house gets water from the same municipal source as he and the rest of the neighborhood. Obviously, the man is delusional.
Your dad says they've lived next to the man for 15 years. He was a perfectly reasonable neighbor, quiet, kept to himself. He and your parents occasionally shared gardening tools and advice. Things were normal until two days ago when your parents were outside weeding and he started shouting at them. It was mostly incomprehensible, but he’s been fixated on poisoned water.
Dad thought he'd maybe had too much to drink and your parents went inside expecting it was the last of it. Yesterday he started pounding on the door, shouting random accusations. “Does he have a history of psychiatric disease or dementia?” you ask.
“We don’t know him that well” your Dad says, “but he was fine for the last 15 years.”
“Any family members?” you ask. Your parents shrug saying, “His life seems pretty solitary.”
The neighbor doesn’t seem particularly dangerous, but these are your parents and you don’t want them to be harassed. Especially not by someone living next door.
“Should we call the police?” You ask. Your dad shakes his head. Leave it alone. He'll stop eventually.”
You brought some pastries for dessert, but left them in the car. You run out to get them. The neighbor is in the middle of the street. With no clothes on, stark naked. He takes one look at you, starts shouting, and moves in your direction with an unsteady gait. This is a situation. It’s not safe for him, your parents or anyone else.
He’s agitated – obviously- and responding to external stimuli. The medical term for hearing voices or seeing things that aren't there. Hallucinations. Trying to gage the situation, you ask him how he’s doing. He screams something incomprehensible. You ask if he wants a pastry. He responds with something about poison.
It'd be nice if he could be reasoned with, and you could take him to the emergency department, but that's clearly not happening. You call 911 and tell them there's a gravely disabled patient in need of medical evaluation.
The medics arrive, and predicably, there's a lot of melodrama. The patient is not amenable to being taken to the emergency department, to put it mildly and half the neighborhood is outside, watching. Things aren’t going to go smoothly, nevertheless, this is the right thing to do. One lady, whose chatted with him while walking her dog, confirms he doesn’t have any family. His parents are no longer living; he never had a long-term partner or any children.
Different states, not to mention different countries, have various laws and regulations applying to patients refusing care. Altered patients, psychiatric patients, etc. In the US, you have the right to refuse medical care and make bad decisions as long as you are sound of mind. If, however, you are disoriented, or a danger to yourself or others, you lose the right to refuse care. Running around naked in the middle of the street poses a clear danger to yourself. So, at this point in time, he doesn’t have the capacity to refuse care. The medics give him a dose of a sedative to calm him down and manage to get him into the back of the ambulance.
You’ve run out of time for dessert. You leave the pastries with your parents and go in for the night shift. You guessed it, the neighbor is now your patient. The nurse calls you into the room, because he's trying to get off the stretcher to leave. He’s extremely agitated, shouting and attempting to punch the security guards.
You go in and try to talk to him again. What’s the point? Number one, to see if we can calm him down and get him to cooperate. I hate sedating patients for both medical and ethical reasons.
Second, we need to see if he really does lack capacity to refuse care. Capacity assessments can become a complicated psychiatric, medical and legal problem. But initially in the ED, it’s a few quick questions. I ask the patients their name, month and year. I used to ask who the president was, but had to stop as it’s became too politically fraught. If they can’t answer the orientation questions, they can’t refuse care. If they can, then the next step is to see if they understand the consequences of refusing care. For example, if you have an undiagnosed infection, do you understand it could get worse, leading to septic shock and death. Are you ok with this risk? If you refuse treatment for diabetes you could lose a leg, lose your vision, go into a coma and die. Are you ok with this? If they are both oriented and can understand the consequences of refusing care, they can leave.
You ask your patient his name, he answers appropriately. You ask the month, he responds with shut up and stop talking to me. You ask him the year, he tries to punch the security guard again. He’s also mumbling to someone who is not there.
What now? That's question number one.
A. chemically restrain him.
B. physically restrain him.
C. check his blood sugar.
Answer: B. Sadly, we need physical restraints. He’s a danger to himself and others, and can’t refuse care. Chemical sedation is safer than physical so it’s a good answer, but it takes time to kick in, especially if you have to give the shot intramuscularly because you can’t get an IV. If you said C, blood sugar, great answer too, we need one. This altered mental status could be a simple as hypoglycemia. It’s not the right answer at this moment, because we need to do something first in order to get blood.
I’d apply physical restraints, temporarily, give him sedation in the hopes of removing the restraints in 30 minutes or so. We could have a long discussion about what to use. If you ask 4 doctors about the best agent for sedation in the elderly, you’ll get at least 6 opinions. In this case a low dose antipsychotic would be my choice.
What is wrong with this guy? At this point, anything is possible. We need a broad workup looking for intracranial hemorrhage, infection, and of course toxins. I’d love to know his baseline mental status. If he has a history of psychiatric disease or dementia this could be a flare. Unfortunately, we don’t have that information as is often the case in the ED.
He remains agitated, but the nurse is able at least to place an IV and get some vital signs. Temperature 98.5F, or 36.9 C. Heart rate 90 bpm, blood pressure 138/70, respiratory rate 20, pulse ox is 100% on room air. After a second dose, you and the resident are able to do a limited physical exam together, given that he's not cooperative. You look for external signs of trauma, particularly on his head, noting no contusions or abrasions. You’ can’t do a cranial nerve exam unless the patient participates, but there’s nothing obvious like a facial droop. His pupils are normal. Heart and lungs are clear, no abdominal tenderness. He’s moving all extremities with surprising strength for an elderly man. He has a rash on his face, it’s red and irritated and looks like a bad case of acne. It doesn't look life-threatening, which is sometimes as far as I can get with rashes.
The nurse is able to obtain blood then removes the physical restraints. She says his blood sugar is 100, normal and sends off his labs. The resident orders an EKG, chest x-ray, and urine specimen to check for infection. He wheels him down for a head CT.
You're in the midst of taking care of a patient with an intracranial hemorrhage after a motorcycle accident when the resident interrupts you. Saying, “sorry can I just tell you about these lab results. I got a call from the lab.” The patient's chloride is 170, normal is around 100 mmol or meq/L. He also tells you the patient has a negative anion gap of 65. A normal anion gap depends on the lab analyzer but is somewhere around + 8-16.
“An elevated gap you say?” Right up the alley of every toxicologist. We've talked about the pneumonic mud piles related to methanol and aspirin toxicity for example.
“No,” he says. A negative anion gap.” You frown at him negative gaps are the opposite of elevated gaps. They are very rare, and don’t have the same implications. The motorcycle victim convulses in a seizure. You tell the resident it doesn’t make any sense, it’s probably a lab error and tell him to resend another specimen.
An hour later, the resident has the repeat labs, this time you can give him your full attention. He says the head CT was negative, no bleeding or masses, as is the chest xr-ay and EKG. The patient hasn’t given a urine specimen yet to rule out a urinary tract infection. But the second set of labs came back exactly the same as the first. Chloride 170, and gap - 65.
Question number two and this is a tough one, more for medical practitioners. What causes a negative anion gap?
A. High Chloride
B. Lithium
C. High magnesium.
D. All of the above.
Answer D. all of the above. The anion gap is an equation measuring the balance of + ions, sodium, against negative ones like chloride and bicarb. The elevated chloride is clearly causing the negative gap in this case. Rather than an acid base disturbance affecting bicarb.
Is this lithium toxicity? There’s no harm in sending a level, but it’s not at the top of my list. It does cause altered mental status but not wild agitation like this and it causes a tremor which he doesn’t have. His magnesium level is normal, anyway high magnesium causes muscle weakness.
Back to the elevated chloride. I'm gonna keep this list brief, but it is kind of interesting. Dehydration, if you lose all of the water in your body, chloride goes up. But his kidney function is normal and he’s not dehydrated on exam. Other ways to lose water is SIADH, syndrome of inappropriate diuretic hormone, where your kidneys lose water when they shouldn't. Severe diarrhea, diuretics ie water pills. Alternatively, it’s extremely hard, but not totally impossible to eat too much chloride, salt for example. I’ve seen most of these things in the ED, I’ve never seen any of them cause more than a very mild and very insignificant change in chloride.
A nephrologist would have some really fascinating thoughts about this electrolyte disturbance. But this isn't a nephrology podcast, it's a toxicology podcast so one thing comes to mind when you see labs like this.
Question. 3 It's time to pick your poison. Is this?
A. bromide toxicity.
B. Chlorine toxicity.
C. Phosgene toxicity
Answer: A. Bromide toxicity has to be on the list if you see an elevated chloride and a narrow gap. Chlorine and phosgene are both often in gas form and weapons of mass destruction used during WWI. They cause respiratory problems and mucus membrane irritation.
What does bromide toxicity cause? Altered mental status and rashes. You can send a bromide level, but it will take at least a week to get back. So we have to go on clinical judgement, first ruling out other causes like urinary tract infection. I’d definitely consult nephrology to see if they have any other ideas. His clinical picture fits with bromism and the classic lab results are an elevated chloride with a negative gap, so it’s pretty convincing. In fact, one of my favorite medical titles is a paper on bromism titled Mind the Gap.
Let’s jump ahead and say we’ve ruled everything else out. The burning question I have now is how he got bromide toxicity. It’s not at all common, at least these days. More on that in a minute.
Question 4. Bromide toxicity causes an increase in chloride in the body.
A. true
B. False
Answer is actually B. It’s tricky question, false despite what I just said. The lab results don’t in fact reflect a truly high chloride concentration in the body. Instead, it’s an interference with the lab test. Many machines falsely measure bromide as chloride. Interestingly, bromide acts like chloride inside your body as well.
How does it cause toxicity? More than acting like chloride, what it does at the cellular level isn’t well understood. We, actually do know a lot about the symptoms. Why? We used to use it a lot in medicine. First, let me clarify the different types of bromine exposures. The symptoms our patient has, bromism, is caused by bromide salts. Methylbromide is a gas used to fumigate houses to get rid of pest infestations. It’s extremely toxic, much more so than bromide salts, but it causes totally different symptoms. Stay tuned for next week when we discuss what happens if you go inside a house being fumigated. We’ll touch on methylbromide exposure then.
Bromism causes a range of symptoms from mild confusion, irritability, and headache, to severely altered mental status with auditory and visual hallucinations, psychosis and coma. Patient often have abdominal complaints, like nausea, or pain. Bromides also cause rash. The classic rash is described as an acne-like rash on the face.
Back to our patient. How do we treat this? Is there an antidote? No. First, and most important, stop the exposure. We still have no idea how this happened, but it shouldn’t continue in the hospital. Second is to try to help the body eliminate bromide by flushing it out with IV fluids. Specifically, you want to use sodium chloride ie normal saline. Why? Because chloride as I said, is similar to bromine inside the body, extra chloride competes with the bromine in the kidneys and increases elimination. Diuretics may help to further increase it.
For our patient, I’d start with several liters of normal saline. Since his kidneys are ok, I’d add gentle diuresis. Whatever you do, this isn’t going to resolve quickly. We don’t know how much he took, so it’s impossible to predict durations, but for sure, bromine has a long half-life, some estimate as long as two weeks. It’s going to take at least a day, probably more. He’ll probably need continued sedation until his mental status starts to improve.
Bromine is removed by dialysis. Should we ask nephrology to dialyze him? This is a tough question and will be a judgement call either way. It’s indicated for severely poisoned patients. What does that mean? There are no exact criteria. On the one hand I do think our patient is severely poisoned. He’s extremely agitated, requiring sedation. On the other, dialysis is an invasive procedure with a lot of side effects and so nephrologist use it judiciously. Additionally, our patient isn’t cooperative and can’t consent to placement of a dialysis catheter. You discuss it with nephrology and agree the best option is normal saline and time to see what happens, with the hope he’ll improve with less invasive, conservative measures.
You call the hospitalist and get admit him to the hospital for ongoing treatment and monitoring. We still have a big unanswered question. Where did it come from?
Let’s talk about the history of bromine use in medicine, then come back to our patient. It has a long history of used in medicine. And until this podcast, I had thought that it was all in the past. Bromine was actually the first antiepileptic, or anti-seizure drug in 1857. It was very widely used as sedative in psychiatry up until the 1950s. Most of what we know about bromism is from old psychiatry literature. Potassium bromide was the most common preparation. It fell out a favor, partly due to side effects, but mostly because it was replaced in the 1950s by barbiturates, which are much more effective and more reliably dosed. It's estimated that between 2 to 8% of all psychiatric admissions were due to bromine toxicity astonishingly. It’s not a huge percentage, but that’s a crazy number of cases of admissions caused by adverse drug effects.
Not only was bromine used it psychiatry. It was available over-the-counter as Bromo-seltzer, marketed for headaches and gastrointestinal upset. It's owner made a fortune and built a 15 story clocktower in Baltimore, still there, called the Bromo-seltzer tower. The original formulation had not one but two toxic ingredients. Bromide and Acetanilide. Acetanilide causes methemoglobinemia. Eventually both were removed.
We talked about how Prozac was frequently mentioned in pop culture last week. This was similar with bromide. It was so common it became a colloquialism, meaning something trite that dulls the mind.
I assumed that bromide was no longer found in Medicine as it's toxicity became widely recognized. In 1975, it was removed from most medicines in the US. But it’s not completely gone. It’s in dextromethorphan or cough syrup as dextromethorphan hydrobromide. Patients develop dextromethorphan toxicity long before they do bromide toxicity. And in fact, it's in medicines I use frequently in the emergency department, for intubation. Rocuronium and vecuronium are actually Rocuronium and vecuronium bromide. Medicine used for myasthenia gravis, pyridostigmine is a bromide, and there's actually a few case reports of patients developing toxicity in high therapeutic doses.
Back to our patient. After day two, he no longer requires antipsychotics, though he remains confused. His chloride level is still extremely high, which the team is using as a proxy measure of his bromide level.
After four days, he's much improved and back to his baseline mental status. Finally he tells you what happened. His primary care doctor recently diagnosed him with hypertension and advised a low salt diet. The patient, a retired IT person, consulted an AI chatbot, about and alternative to salt, sodium chloride. The chatbot recommended sodium bromide. Where did he obtain it? Online of course. He ordered Dead Sea Salt. It was supposed to be used as a bath salt, not ingested because it’s well known to be toxic when eaten. Soaking is fine, I’m sure you know you can swim in the Dead Sea as long as you don’t swallow the water.
Our patient, unfortunately, based on the dangerous advice of the chatbot, decided to replace sodium chloride in his salt shaker and food with sodium bromide. He ate the Dead Sea salt meant for bathing in. His bromide level eventually returns at 2000 mg/dL, massively elevated. There's a wide range of levels considered toxic, anywhere from 400 mg/dL up to 1500 mg/dL considered toxic levels. You explain what happened, He promises never to eat dead sea salt again.
This case is fictional as are all our cases to protect the innocent. It is based on real cases of ingestion of dead sea salt, as well as a very disturbing real case of a man who decided to replace sodium chloride with sodium bromide, apparently based on the misguided advice of an AI chat bot. I hope you take this as a cautionary tale. While AI is an amazing resource and tool, we physicians love to use it, its medical advice is far from 100% reliable. There are reports it may have worsened medical conditions like psychosis and suicidality.
That brings us to the last question in today’s podcast. I mentioned the toxic ingredients in bromo-seltzer. Acetanilide was replaced with acetaminophen or paracetamol. There’s an interesting parallel with the Tylenol murders in the 1980s with poisoned Bromo-seltzer 100 years earlier in 1898. A man was sent a package of Bromo-Seltzer. His aunt took some for a headache and died shortly thereafter. It wasn’t bromide toxicity. Her doctor took a small sip from the bottle - I don't recommend this diagnostic approach by the way- and collapsed. He was nearly the second victim, but he survived. Both Tylenol in 1980s and Bromo-Seltzer in 1890s were poisoned with what?
A. Strychnine
B. Arsenic
C. Cyanide
D. Thallium
Follow the Twitter and Instagram feeds both @pickpoison1 for the answer. Remember, never try anything on this podcast at home or anywhere else.
Thanks for listening. It helps if you subscribe, leave reviews and/or tell your friends. Transcripts are available at pickpoison.com.
While I’m a real doctor this podcast is fictional, meant for entertainment and educational purposes, not medical advice. If you have a medical problem, please see your primary care practitioner. Until next time, take care and stay safe.