Analeptics – Quinn’s Toxicology Post

Analeptics: The Modern Coma Cocktail


I had never actually heard this term analeptic before and it seems somewhat old fashioned.

It means something that is restorative and invigorating, in this context, used to wake up someone with a decreased level of consciousness.

First, a quick practice case:

A middle-aged man presents to your hospital obtunded with poor respiratory effort. Which of the below treatments/Medications do you administer?

  1. Place patient in Ice bath
  2. Strychnine
  3. Camphor.
  4. Picrotoxin
  5. Caffeine

Of course all of these answers seem ridiculous, if not downright dangerous. However, depending on the where and when of your medical practice, any one of these may have been the standard of care.

At one time or another, these have been treatments of choice for decreased levels of consciousness. And some of them even work… In a fashion.  Strychnine for example might even wake you up so hard that you are totally conscious for the painful muscle convulsions that may eventually cause you do die of anoxia. Ew. But at least you woke up, so that’s good.

Granted, these are from a time where mechanical ventilation was not really a thing, so you would go to great lengths to avoid deep coma. No matter the cost.

The risk benefit here is a little different than most things we deal with nowadays and was basically “probably dead vs possibly dead”

Such is the dubious history of Analeptics and the “coma cocktail,” which is still evolving and is argued about to this day.

The coma cocktail:

In the modern era, say for the past 20 years or so, this contains dextrose, thiamine, naloxone, and flumazenil.

The concept for this when it was first devised is that it was something you could just reflexively give to people with decreased levels of responsiveness, to rapidly address reversible causes of depressed consciousness while reducing permanent neurological injury.

You are probably quite familiar with the medications listed above, although you may not use all of them, and may not use any of them in the reflexive manner that this concept was designed for.

The prevailing practice attitudes now have shifted against indiscriminate administration of this cluster of potentially harmful medicines, so it is probably most appropriate to break it down and look at each component individually.


Dextrose is a sugar, used to treat hypoglycemia.

Hypoglycemia is a common endpoint that may result from a drug or toxin exposure, nutritional deprivation, or numerous nontoxic events.

Hypoglycemia is common. The rate of hypoglycemia in patients with altered mental status of any cause is approximately 8.5%. This means that about 1 in 12 of your altered mental status patients have a hypoglycemic component. This is extremely significant because prolonged hypoglycemia can result in permanent brain damage, and mortality in hypoglycemic patients ranges from 11-27%.

This alone might be enough to persuade you administrate this in your altered patients. You would use D50 in adults, and D25 or D10 for pediatric patients.

In the end, the dosage that is required is “enough” because every patient is different, and have their own thresholds at which they become symptomatic. Dosing for pediatric patients is as follows. D10W: 2.5ml/kg,    D25 4ml/kg. In adults, starting with 1-2 ampules of D50 is appropriate, each ampule contains 25g of dextrose.

What if they get too much?

Concerns for making the blood sugar too high are generally not demonstrated clinically.  Renal losses alone keep it from going way too high, which is something we routinely take advantage of in patients with DKA and HHS. And anyway, one ampule of D50 only generally raises blood glucose by 60mg/dL.

What if it wasn’t hypoglycemia at all?

If their altered mentation is from something nontoxic, like an ICH, there has been no demonstrated long term change in survival or functional ability.

The solution to these concerns is to be able to reliably detect hypoglycemia right?

Well, if we accept that treatment should be administered as fast as possible to reduce hypoglycemia, then bedside testing might be completely impractical.

On the flip side, and in defense of standard bedside testing, a potentially very high number of patients (up to 25!) are misdiagnosed based on purely clinical findings, as many appear agitated instead of somnolent, and for this reason their blood glucose may not be tested, so maybe we should just test everyone.

Dangers of relying on a number:

Keep in mind though that lots of people become altered at blood levels above the “definition” of hypoglycemia of a glucose below 60. If a person lives their life at a serum level of 300mg/dL, the normal lab values become less useful.

This justifiably seems like a tough balance. The clinical need to make a rapid decision in a busy or hostile environment vs a documentable objective measurement.

Hoffman and Goldfrank recommend a stratification system for your altered pts:

  1. If they have numerical hypoglycemia, treat. Problem solved.
  2. If they have nonfocal neuro exams, and have borderline glucose levels. Give dextrose.
  3. If you don’t have ability to test a patient in the above scenario, give dextrose anyway.
  4. So what do you do if they are altered and have a localizing exam: This is admittedly rare in cases caused by low sugar, 2.5%. So for these patients, if you go strictly by the numbers, you will get 90% of those rare patients. And even if you are wrong, the extra sugar won’t hurt.

So, with this method, you will admittedly over treat a large number of patients, but you probably won’t delay a single person with real hypoglycemia. Which is important, because the cost of an amp of D50 is like 5$. But the economical and physical cost of not treating it is massive.


Right off the bat, empirical use seems less complicated than with dextrose.

Thiamine (Vitamin B1) functions as a cofactor for pyruvate dehydrogenase, which links anaerobic glycolysis to the krebs cycle, as well as an enzyme in the krebs cycle itself, and the pentose phosphate pathway. It’s important.

One interesting, and possibly problematic aspect of thiamine is that the amount needed is dependent on total energy intake. More energy=more thiamine needed.

Dietary thiamine becomes decreased  in chronic liver disease, folate deficiency, malabsorption, malnutrition like with TPN/post op patients, or those juice cleanses your facebook friends from high school like to use and/or sell.

The prime example of people who take in calories without other nutrients (like thiamine!) are alcoholics. Liquor has lots of energy but not so many vitamins. The most notable result of low thiamine is everyone’s old friend Wernicke-Korsakoff.

The characteristics of Wernicke encephalopathy includes oculomotor abnormality, ataxia, and confusion. It carries a mortality rate of 10-20%, and up to 80% of survivors develop the other half of the odd couple- Korsakoff psychosis with that fun, permanent short term memory damage. Confabulation is all fun and games until you need a reliable medical history from a patient besides a lifetime of drinking.

So, this is rare. But it is very bad.

Including the 100mg IV of thiamine treats/prevents the encephalopathy. And it costs like 1$. That’s what we in the biz like to call a steal.

But what about that thing with glucose I learned about in medical school that I also had to know for step 1 and 2!?!

And of course the other benefit you may hear, is that it prevents the precipitation of the encephalopathy by the dextrose loading these people will get too, as thiamine requirement is calorie dependent.

So that part sounds right and almost makes sense from a basic physiology standpoint, but the evidence for this is very lacking. The cases that drive this are most likely people who had the encephalopathy already, or who got dextrose in large volumes for a long time (hours to days) without other nutrition.

Withholding dextrose until thiamine doesn’t do much. Thiamine uptake is way slower than glucose anyway, something like 6 hours.  So even giving them together wouldn’t even help that much if you were truly worried about Wernicke’s. Which again, you shouldn’t be.

So in that case you would basically just be withholding dextrose from a patient that may desperately need it to avoid something that is probably theoretical at best. Don’t be that person.

Now you have me all fired up to give thiamine!

You can give it oral, IM, or IV. People who need the coma cocktail probably won’t be getting it orally if they are really that altered. At least they shouldn’t. Most medications are less effective if aspirated. And if they are really malnourished, they might not have good muscle mass to take an IM injection. You can do it that way, but you will have an IV in these people anyway.

Thiamine is cheap, safe, and since you can’t measure it in a patient, this has a good place in the modern coma cocktail. At worst, it just reminds us to address nutrition, right?


Probably the best known and most talked about. It is a pure opioid antagonist used for reversal of acute intoxication. It rapidly counteracts sedation, respiratory depression, miosis, analgesia, bradycardia, and GI stasis caused by exogenous opioids through your 3 different opioid receptors.

It also fights your endogenous opioid peptides too. Which may explain why it works a certain degree on some intoxications with things that are not opioids, such as some antiepileptic medications.

Naloxone can be given IV, IM, through an ET tube, and intranasally.

There are risks like pulmonary edema and hypertension, and risk to you through precipitation of violence. However, these risks to the patient are generally considered rare. Approx 1% or less.

Except for extremely unpleasant opioid withdrawal. We know it does that. That’s the whole point. You can do that every time if you really wanted to. The effects of withdrawal are not life threatening, but it can cause other problems to arise in the setting of a polydrug intoxication. You don’t want your newly opiod-free patient vomiting from withdrawal while their benzodiazepine co-ingestion is keeping them nice and asleep.

On the other side of the same coin, you could uncover dangerous and more threatening sympathomimetic symptoms from something like cocaine, like seizure and arrhythmia.

Are there any agitated people this might help with? Like dextrose does?

This differs from the case of dextrose possibly helping agitated people. People on opioids won’t be rowdy like that. You aren’t gonna help, and will probably make things worse.

What is the best metric for giving this?

This is something you want to consider giving right away to someone with decreased consciousness. In the Hoffman/Goldfrank paper, they reference prior works by Hoffman that show a respiratory rate of 12 of less predicted a naloxone responder 80% of the time. So this is where you want to focus your attention if you are thinking about opiods. Respiratory depression is what kills these people, so treat the respiratory depression. A PCO2 from a blood gas is better, but that takes a lot more time than checking out the respiratory rate.

But the PUPILS. People on opioids have miosis!

But what if they aren’t miotic? Certain opioids have a much more dramatic effect on respiration than pupils. They could die of anoxia with normal pupils. Our old friend Demerol was particular well known for that. Pupils are nice, but you can’t breathe through your pupils, to quote Dr. King.

So how much do you use?

Conventional dosing ranges from .4 to 2mg. If you reach 10mg, isolated opioid toxicity is unlikely.

The Hoffman paper suggests that you lower the dose to .2mg as a first dose. That was in 1995.

This new publication in the British Journal of Pharmacology suggest that even that may be unnecessarily high, and that a first dose of 40 micrograms is more appropriate. This way you can titrate to their breathing, which is what the lethal issue, not the sleepiness. They can be sleepy and breathe, as long as they are protecting their airway.

And what if they DO need the full 2mg to breathe? Naloxone acts so fast that you can titrate it up .04mg at a time, and it won’t keep you in the resuscitation bay for an hour. It will still be fast. They can be bagged or intubated during this of course.

Go slow. Don’t hit them with 2mg right away. If you need it, work up to it. You avoid getting punched and pooped on, and the patient doesn’t leave AMA only to have their naloxone wear off 3 blocks away.

The recommendations for naloxone drip remain the same, with an hourly infusion rate around half of what you needed to give them initially.


Last and probably least, is flumazenil. While naloxone got a WHO classification as an essential medicine, flumazenil got a black box warning.

This is no longer a standard, and essentially only exists in the coma cocktail as a historical item.

It reverses sedation from benzodiazepines, but will also precipitate withdrawl in patients dependent on them, or alcoholics.

Compared to opioids, this is less on the gross side, and more on the potentially fatal side. It can precipitate dangerous withdrawal seizures which you would then treat with benzos… which you just nullified the effectiveness of.

To be fair, it was developed for conscious sedation reversal before it was added to the coma cocktail. Which it still has a valid use for- If you have a clear history, and a low suspicion of benzo/ETOH dependence it may have a role in reversal of conscious sedation from benzos. OR… you could just intubate them and let them get better on their own.

If you are curious, Initial dosing is .2mg for adults, slowly given/titrated. In kids it is 10ug per kilogram with a max of .2mg.

You essentially want to give this the same time you perform gastric lavage. AKA basically never.

Probably best to avoid without advice from a toxicologist



  1. Hoffman, Robert S. “The Poisoned Patient With Altered Consciousness.” Jama 274.7 (1995): 562. Web
  2. Sivilotti, Marco L.a. “Flumazenil, Naloxone and the ‘coma Cocktail’.” British Journal of Clinical Pharmacology Br J Clin Pharmacol 81.3 (2015): 428-36. Web.




Leave a Reply