Antipsychotics Don't Help ICU delirium

by Drew Rosielle (@drosielle)

It just gets worse and worse for the idea that antipsychotics have efficacy for delirium.

Last year I posted about the RCT of haloperidol, risperidone, or placebo for delirium symptoms in 'palliative' patients. I'm pretty sure I called for more controlled, 'high quality,' trials, and we are lucky enough to have another.

This one is a randomized, double-blinded, registered, controlled trial of haloperidol, ziprasidone, or placebo for ICU delirium, just published in NEJM.

The trial took place in a geographically diverse group of US-based intensive care units. They enrolled adult patients in medical or surgical ICUs, who had delirium detected by the CAM-ICU (a well-validated tool: lots of info on it here if you want to read more). They did something interesting - they consented patients for the study (some of them at least) before the onset of delirium.  They also did prospective, broad case-finding, using research personnel to evaluate patients twice daily to see if delirium was present or not. They did not rely on clinical personnel in the ICU to identify delirium but instead proactively found them, which is a nice touch given how under-recognized delirium can be. My sense is that they consented patients rapidly upon admission to the ICU (or their surrogate), then followed them closely to rapidly identify those who developed delirium.

Once delirium was identified, patients were randomized to 2.5mg of haloperidol or 5mg of ziprasidone (or matching placebo) IV, then q12h. (Older patients received lower starting doses.) The dose of the study drug/placebo was doubled with each subsequent dose if the patient did not improve up to a maximum of 20 mg a day of haloperidol or 40 mg a day of ziprasidone. They halved the dose of the drug if a patient later had no delirium per the CAM-ICU, or stopped it if a patient was on the lowest dose already, or if the patient had 4 consecutive assessments indicating no delirium. Regardless, the trial drug/placebo was stopped at 14 days or ICU discharge if the patient made it that far and was still delirious.

The primary endpoint was days alive without delirium (per the CAM-ICU) or coma.

566 patients underwent randomization, 89% of those had hypoactive delirium. Median age was around 60 years old, 43% female, ~83% were white.

The median duration of exposure to a trial drug/placebo was 4 days, with the mean daily doses of 11 mg of haloperidol and 20 mg of ziprasidone.

Active treatment did not seem to do anything. The median number of days alive without delirium or coma was the same in both groups (8.5 vs 8.7). Secondary outcomes were similar between 'active' and placebo groups too - 30 and 90-day survival, time to discontinuation of ventilation, time to ICU discharge, ICU readmission, time to hospital discharge -- all the same.

Safety endpoints were the same across groups too with the exception that QT prolongation was more common in the ziprasidone group (it's not clear to me if this was statistically significant or not). Other side effects, including excessive sedation, were no different between placebo and the study interventions.



What does all this mean?
At the end of the day, I think it underlines the idea that antipsychotics have no routine role in the management of delirium.

It took me a while to accept this, but at this point I'm looking for any compelling data to suggest otherwise. We just don't have it, and I've mostly stopped prescribing antipsychotics for delirium. While any study like this has elements you can pick apart (e.g., I wouldn't exactly dose haloperidol like they did), at the end of the day they used reasonable doses, doses that I have commonly used, and they allowed relatively high doses to be used (this was not a study of super careful, wee doses). They identified the patients prospectively and rigorously - I don't think it likely that we're going to see a substantially better-designed study in the near future than this.

Notably, most of these patients had hypoactive delirium, and they did not break down their findings by patients who were hyperactive vs hypoactive. I imagine some people are going to argue that they should have done that but overall I'm glad they didn't. It was only 10% of their subjects, and it's better they didn't data-mine their findings that way. It could have been a prespecified analysis, which as far as I can tell it was not, and the study wasn't powered to detect differences between so few subjects, and so I'm glad they left it be.

I think it would be reasonable to argue that we still really need a similarly well-done, large, study only for hyperactive ICU delirium - I think we do - but I also don't see any reason any of us should advocate for the idea that antipsychotics are effective for hyperactive delirium in the meantime. They might be, but I don't think that's very likely personally. If someone does it, I really hope they include quetiapine as one of the arms, as it's used a lot around many institutions, and I suspect for the worse (although I truly don't know), and would welcome any good data on it.

At this point, I think a lot of the perceived benefits people have of antipsychotics is that they sometimes sedate people, even if only mildly at times, and so patients 'look better,' but we've actually done nothing to modify the course of their delirium or improve their cognition.

Sedating someone can sometimes be a perfectly reasonable/helpful/legitimate therapeutic outcome: for very agitated patients who may pose a risk to themselves, or for patients very near the end of life for whom our care goals are really centered on preventing and actively minimizing any sources of suffering.

But it's not a great outcome for most of our other patients, and like a lot of things we deal with (eg, cancer-associated appetite disturbances and muscle loss) delirium is a complex phenomenon and needs complex interventions to prevent/mitigate, and perhaps the entire idea that there is a 'right drug' (or drug class) to help it was wrong all along.

For more articles by Drew Rosielle, click here. For more articles about delirium, click here. For more articles about the ICU, click here.

Drew Rosielle, MD is a palliative care physician at the University of Minnesota Health in Minnesota. He founded Pallimed in 2005. You can occasionally find him on Twitter at @drosielle. 

Reference
1. Girard TD et al. Haloperidol and Ziprasidone for Treatment of Delirium in Critical Illness. N Engl J Med. 2018 Oct 22. doi: 10.1056/NEJMoa1808217.