The choice of sedation in critical care patients who are mechanically ventilated may be independently associated with poor outcomes, including increased mortality rates, according to the results of a new study.

The study found that patients sedated with propofol spent less time on the mechanical ventilator, one day less in the ICU and four days less in the hospital than patients sedated with either continuous or intermittent benzodiazepines. What surprised the research team, however, was the finding that patients sedated with benzodiazepines had statistically significant increases in both ICU and hospital mortality rates compared with the propofol group.

Indeed, ICU mortality rates were 18.4% in the propofol group, compared with 26% in the continuous benzodiazepine group and 22.7% in the intermittent benzodiazepine group (P<0.001). Hospital mortality rate was 28% for the propofol group, 36% for continuous benzodiazepine and 35% for the intermittent benzodiazepine group (P<0.001).

“I was really caught off guard by the mortality numbers,” said Nick W. Lonardo, PharmD, clinical pharmacy coordinator for the surgical ICU at the University of Utah Hospital and Clinics in Salt Lake City. “We are finding that these drugs are independently associated with worse outcomes.” The researchers reported their findings at the 2009 annual meeting of the Society of Critical Care Medicine (SCCM), in Nashville, Tenn. (abstract 674).

Dr. Lonardo’s group reviewed data from the critical care database Project IMPACT to study if the choice of sedation was associated with increased ventilator times, hospital stays and mortality rates. The investigators narrowed their search to patients with single ventilator events, who received only one sedation agent and who were mechanically ventilated for at least 48 hours. The database identified 21,872 patients, from 131 ICUs, who matched the study entry criteria.

The study found that patients receiving continuous or intermittent benzodiazepines spent significantly more time on a ventilator than did those who received propofol (8.1 and 7.8 days vs. 6.8 days, respectively). In addition, the percentage of patients on a ventilator for more than one week was significantly higher in the two groups receiving benzodiazepines (39%) compared with patients on propofol (30%). As noted, the length of stay in the ICU and hospital were also longer in the benzodiazepine groups. The difference between one extra day spent in the ICU and four days in the hospital is significant in terms of overall patient outcome and hospital costs, Dr. Lonardo noted.

Choice of Sedation

Although surveys have reported that midazolam is the most commonly used sedative in the ICU, this was not the case in Dr. Lonardo’s study. The vast majority of patients were sedated using propofol (n=15,763), followed by continuous benzodiazepine (n=5,366)—either lorazepam or midazolam—and intermittent benzodiazepine (n=743).

“What our study showed was that about 72% of patients were sedated with propofol,” Dr. Lonardo said. “In addition, the percentage of patients treated with propofol did not change between 2001 to 2007 despite the release of SCCM guidelines on sedation in 2002 that recommended the use of either continuous benzodiazepines or intermittent benzodiazepines as the preferred agents for most patients.” In fact, his own unit switched to propofol almost six years ago.

Although Dr. Lonardo does not know exactly why benzodiazepines were associated with increased mortality in the study, he speculates it may be linked to increased delirium among patients on benzodiazepines. Wesley Ely, MD, a critical care expert at Vanderbilt University in Nashville, and colleagues have shown that delirium is an independent predictor of higher six-month mortality and longer hospital stays, Dr. Lonardo said (JAMA 2004;291:1753-1762). In Dr. Ely’s study, lorazepam was also identified as an independent risk factor for delirium.

“My speculation is that benzodiazepines could be inducing a more sustained delirium than propofol and, therefore, causing higher mortality.” But, he added, the study was purely an association at this point.

Evidence continues to mount, however, for a change in sedation policy. “Based on the work done by [Dr. Ely’s group], we now know that delirium is not benign,” Dr. Lonardo said. The era of keeping patients highly sedated or overly sedated is over. “We are now learning that it is not necessary, and it can have serious consequences.”

Simon W. Lam, PharmD, clinical pharmacy specialist at the Cleveland Clinic Foundation in Ohio, said the link between delirium and worse outcomes is well known. “We have also known that benzodiazepines are associated with delirium. Some people have made the leap that benzodiazepines are associated with higher mortality because of the increase in delirium. Thus far, to my knowledge, there has not been a paper that has proven that.” Therefore, if this is the first study to prove that an agent that is associated with less delirium improves mortality rates in the ICU and hospital, “this would be very interesting.”

John Lewin, PharmD, clinical specialist at The Johns Hopkins Hospital Department of Pharmacy and School of Medicine in Baltimore, agreed. “We know that the presence of delirium is an independent predictor for a worse outcome. What we don’t know is whether or not outcomes can be improved by effectively treating or preventing delirium. This seems to be the next step for further investigation.”