Delirium tremens

The evidence level at which psychiatric intensive care works is quite depressing. We tried to collect the data on how to treat delirium tremens, but apart from some small case series and review articles such as this recent one, there really is not much to go on.

So here are some of my experiences:

  • Make sure to strictly distinguish delirium from withdrawal. There is a lot of difference in terms of physiology as well as treatment. And delirium is treated in a monitoring unit, withdrawal can be done everywhere. I am not sure whether good withdrawal treatment can actually prevent delirium, but we surely try to…
  • Phenomenologically you can separate a) Agitation b) Vegetative c) Psychotic features. These might be linked to separate neurotransmitters in overdrive (a) GABA b) NA c) dopamine), but again this has not been proven.
  • You can treat agitation by filling those GABA-receptors with benzos, barbiturates or modify them with proposal.
  • Vegetative symptoms are best treated with clonidine or dexmedetomidine.
  • Psychotic features usually respond to dopamine antagonists, where the best (iv available) is haloperidol. 
  • As for management, I recommend filling up with diazepam, because of its superb pharmacology, namely it’s accumulation which leads to slow secondary withdrawal symptoms.
  • The problem is that diazepam has different pharmacokinetics every day and in every patient, so that you titrate your loading dose according to some agitation/sedation scale and must not be scared of high doses (such as 100-400mg). Take your time and every opportunity to see your patient should be used to push some more, first in 10 mg steps, then escalating boluses to 20 or even 30 mgs.
  • Once filled, I halve the dose every day in tid-dosing scheme, e.g. 120 mg on the first day leads to 3×20 the second.
  • If properly sedated (drowsy/sleeping, but arousable on command) and still some vegetative action (in particular heart rate), titrate clonidine in steps of 75micg.
  • If, when aroused, the patient hallucinates or speaks with the wall, add haloperidol (3 x 2,5 to 3 x 5mg).
  • Take care of all organ systems during this trip – do prophylaxis as much as you can. Don’t forget ulcer prophylaxis, mobilization, thromboprophylaxis.




Yet another delirium

Your stroke patient changes his mental status. As for instance these 2 recent studies (Neurology 2011, Stroke 2012) tell us, delirium is quite frequent in stroke, yet it is not clear (to me) whether it is more or less frequent in stroke than in other hospitalized patients, of which about 30% might develop altered mental status, aka delirium.

The german terminology tends to equate “delirium” with the american terminus “hyperactive delirium”, i.e. altered mental status with vegetative and psychomotor activation, leaving the “hypoactive delirium” without a proper label – perhaps this explains the divergent epidemiologic data.

So here is my algorithm for altered mental status:

History (A4D5): The following strongly predispose to delirium

  • Alcohol
  • Angst
  • Atemnot (dyspnea)
  • Aua (pain)
  • Deprivation (sensory deprivation, no hearing/visual aids, no contact)
  • Dehydration
  • Depression
  • Degeneration (dementia and other degenerative disease than Alzheimer’s)
  • Drugs (a long list, all psychotropic, see below)

Labs: SCHNELL Vitamines + Tox

  • Sepsis (CRP, Leukos, PCT, Urine, chest xray)
  • Cardiac (cTnI – MIs can lead to delirium curiously)
  • Heme (anemia)
  • Nephro (Uremia)
  • Endocrine (TSH, Elytes – Addison’s/Cushing)
  • Liver
  • eLytes
  • Vitamins
  • Toxicology

CT/MRI: Stroke, SAE, Tumor etc. (in the Stroke Unit we usually have such images ready, but think of neurologic complications)

EEG: r/o Status and distinguish between withdrawal (fast EEG) and the rest (slowing, metabolic)

LP: r/o encephalitis

As for management,

  • I can only stress the importance of reducing the triggers, in particular psychotropic drugs – remember that opiates are rarely delirogenic, yet Metamizol is (it is one of the most lipophilic of NSAIDs).
  • Withdrawal has to be dealt with.

Think of the non-pharmacologic treatment:

  • get relatives to the bed to hold hands and soothe the patient, if not possible, do it yourself.
  • Provide hearing aids, glasses, a warm and friendly environment without too much but also without too few cues – beware of sensory deprivation.
  • Mobilize, if possible in the corridor, so that they see and hear people.
  • If seeing and hearing is difficult, use a pacifier to provide at least oral sensory stimuli.
  • Give them something to do, to work on.
  • Reorientation: Have calenders and clocks everywhere. Tell them where they are and what they are here for.
  • Reduce physical restraints as much as possible.

What about restraints and drugs? There is not much data about the use of either in the treatment of delirium. I think of most of the therapy as self defense – if they are harming you or themselves, you need to protect them without harming too much.

  • Haloperidol is the standard except for extrapyramidal degenerative patients where it triggers hypoactive delirium. Has few side effects, can be given iv. Risperidone is just like Haloperidol only younger, so more data for delirium in dementia.
  • Quetiapine is good for sedating without too many side effects (except orthostatic hypotension), helps resynchronize day/night, but start slow.
  • Olanzapine is good for rough patients, can be given im.
  • Cholinesterase inhibitors might be helpful, but to do a proper study you would have to define the subgroups that really profit. IMHO it is the sensory deprived semi-demented patients that are kicked over the edge by too much information for a too slow braing that benefit.
  • I personally do like Circadin (retarded L-Melatonin) to resynchronize the diurnal rhythm, but there is obviously no study for it.
  • Please refrain from sedating typical neuroleptics, antihistaminics and antidepressants because of their usually quite pronounced anticholinergic properties [it’s not fair to pull the handbrake during the race].


  • As usual, uptodate has a good article on the topic.
  • Use a pharmacology book to study receptor profiles of neuroleptics.