Patient Safety Tip of the Week


February 10, 2009     Sedation in the ICU: The Dexmedetomidine Study




Those of you who have read our prior columns pertaining to preventing delirium and managing delirium know that use of sedating agents is one of the biggest contributors to the development of delirium in the hospitalized patient. Their use is often a necessary evil in the intubated, mechanically ventilated patient in the ICU, a major reason why the prevalence of delirium is so high in ICU patients.



The most significant sedating agents used in the ICU have historically been midazolam, lorazepam, and propofol. All agree that finding an agent that is less likely to precipitate delirium would be desirable. The drug dexmedetomidine, an α2-adrenoreceptor agonist, has often been mentioned as a potentially useful drug to fulfill that role (Pun 2007, Pandharipande 2007).



Now a new study published in JAMA (Riker 2009) touts the use of dexmedetomidine for sedation in the ICU. That study concluded that dexmedetomidine was as effective as midazolam at keeping patients in the desired sedation range and was associated with a reduced prevalence of delirium and reduced time to extubation. The dexmedetomidine was associated with more episodes of bradycardia, whereas the midazolam was associated with more tachycardia and hypertension. However, the side effects were relatively mild and the overall safety profile of dexmedetomidine was reasonable.



But let’s look more closely at the study (this will also give you insight into how research studies need to be interpreted). First, the study was sponsored by the pharmaceutical company that makes the study drug and most of the authors have received some sort of financial support from that company. Note that when we read a paper we do not look at the “conflict of interest” or “financial disclosures” section until we’ve finished the paper so that we will not be biased in our assessment. However, this paper had the usual tell-tale warning of an industry-sponsored paper in its abstract and discussion: the study failed to meet the primary outcome yet is presented as a very positive study.



The primary outcome was the time a patient spent in the desired sedation range, as measured by RASS (Richmond Agitation and Sedation Scale) scores. The study showed no difference in this primary outcome measure between the group receiving dexmedetomidine and the group receiving the comparator drug, midazolam. (Note that if we were designing this study that would not have been the primary outcome we’d have been interested in anyway. The much more important outcomes to assess are mortality, duration of ventilation, ICU and total hospital length of stay, and overall costs).



Secondary outcomes included prevalence and duration of delirium, use of fentanyl and open-label midazolam, a bedside nursing assessment, duration of mechanical ventilation, and ICU length of stay.



Delirium was assessed daily using the CAM-ICU tool that we previously described in our October 14, 2008 Patient Safety Tip of the Week “Managing Delirium”, though a rather complex statistical analysis incorporating a “generalized estimating equation” was used to “model” the prevalence of delirium. Using that method the prevalence of delirium was 24.9% reduced in the dexmedetomidine group (statistically significant).



There was no difference in the ICU length of stay but the median time to extubation was 1.9 days shorter in the dexmedetomidine group (statistically significant). Our antennas are always raised when we see a study that reports “median” and does not also include “mean” since that often means the data were presented in the most favorable light.



The composite nursing assessment score showed statistically significant better scores for the dexmedetomidine group for 2 of the components (communication effectiveness and cooperation) but this appears to be a tool developed for this study and the authors include no references to the validation of that tool.



There are several other methodological problems with the study. The results were not reported by “intention to treat’ analysis, which is the customary format for studies of this nature. Also, there is no report on the efficacy of the “blinding” in the study. The latter is significant in this case because the side effect profile of the two agents is so different (one causing bradycardia, the other tachycardia) that one might anticipate the blinding was less than ideal. The characteristics of the populations at the time of randomization appear to be similar. They did look at incidence of delirium at time of randomization and it was similar in the two groups but they did not report the incidence of pre-existing dementia in the two groups, which is perhaps the best predictor of delirium. An unequal number of dementia patients could sway the results toward one group. And patients could be entered into the study up to 96 hours of mechanical ventilation before randomization. Probably a more appropriate protocol would have required randomization at the time of intubation and onset of mechanical ventilation.



Also, there was no significant impact on clinically important outcomes like mortality, ICU length of stay, or incidence of VAP (ventilator-associated pneumonia) though the study may not have been powered to assess all of these. And the paper did not report any data on costs of care.



Are there good aspects to this study? Yes. The sedation protocols with daily arousal assessments and use of a standardized tool (the RASS) were excellent. The use of the CAM-ICU tool for delirium assessment is also something that all ICU’s should be doing.



So what do we learn from this study? We’d put it in the category of a “pilot” study that suggests dexmedetomidine may be administered safely and may eventually prove to be a better alternative than benzodiazepines or other sedating agents currently used in ICU’s. But it really just posits that hypothesis. The next step should be a randomized controlled trial looking at use of this drug vs. one or more comparator sedating agents used early in mechanically ventilated patients. However, it must focus on clinical relevant outcomes and correct all the methodological problems we’ve noted above.



Will such a study get done? The FDA has not approved use of dexmedetomidine in the manner in which it was used in this study. However, the FDA has also indicated its intent to allow pharmaceutical representatives to use published papers about “off label” uses of drugs in their presentations to physicians. Our bet is that the latter will significantly increase the use of dexmedetomidine in ICU’s, removing the sole incentive that the pharmaceutical company would have to do the study correctly.



See also our June 23, 2009 Patient Safety Tip of the Week “More on Delirium in the ICU” for further discussion on sedation in the ICU.






Pun BT. Ely EW. The importance of diagnosing and managing ICU delirium. [Review] [90 refs] Chest 2007; 132(2):624-636



Pandharipande PP, Pun B, Herr DL et al. Effect of Sedation With Dexmedetomidine vs Lorazepam on Acute Brain Dysfunction in Mechanically Ventilated Patients: The MENDS Randomized Controlled Trial. JAMA 2007; 298(22):2644-2653



Riker RR, Shehabi Y, Bokesch PM, et al for the SEDCOM (Safety and Efficacy of Dexmedetomidine Compared With Midazolam) Study Group. Dexmedetomidine vs Midazolam for Sedation of Critically Ill Patients. A Randomized Trial. JAMA. 2009; 301(5):489-499. Published online February 2, 2009








Patient Safety Tip of the Week Archive


What’s New in the Patient Safety World Archive