How to Use Less Sedation and Improve Outcomes in Mechanically Ventilated Patients

Abstract & Commentary

By David J. Pierson, MD, Editor

Financial Disclosure: Critical Care Alert’s editor, David J. Pierson, MD, nurse planner Leslie A. Hoffman, PhD, RN, peer reviewer William Thompson, MD, executive editor Leslie Coplin, and managing editor Neill Kimball report no financial relationships relevant to this field of study.

SYNOPSIS: This was a comprehensive, interdisciplinary, 2-year quality improvement project to reduce continuous sedation infusions and improve the recognition and prevention of delirium in patients with acute lung injury. It resulted in less infusion use, more days per patient without sedation, and more patient days awake and not delirious, although the median proportion of days with delirium per patient actually increased.

SOURCE: Hager DN, et al. Reducing deep sedation and delirium in acute lung injury patients: A quality improvement project. Crit Care Med 2013;41:1435-1442.

Hager and colleagues report on a quality improvement (QI) project aimed at decreasing the use of continuous sedative infusions in mechanically ventilated patients, and also at increasing awareness, recognition, prevention, and management of delirium in such patients. To maximize the comparability of the patients and the care they received apart from this intervention in their before- and after-study design, they included only adults with acute lung injury (ALI) managed in a single unit, the medical ICU at Johns Hopkins Hospital. In this unit, case definitions, patient demographics, admission patterns, staffing, and most aspects of patient management (including other protocols) did not change during the 6.5-year study period. After a 30-month control period in which all ALI patients were prospectively enrolled, the investigators carried out a comprehensive, 2-year transition program to elicit staff buy-in and adoption of a new sedation protocol and delirium screening approach, after which they again enrolled all ALI patients for a 20-month post-intervention study period.

Severity of illness, management of mechanical ventilation, and other features of the patients and their care did not vary between the two study periods, according to the comparisons used by the authors. There were 120 patients in the historical control group and 82 patients in the post-intervention group. After implementation of the new sedation protocol, the median proportions of ICU days on which patients received continuous sedative infusions were significantly less: 33% vs 74%, and 22% vs 70%, respectively, for narcotics and benzodiazepines, both P < 0.001. In the second period, the median Richmond Agitation Sedation Scale (RASS) score was -1.5, compared with -4.0 in the control period, indicating greater wakefulness (P < 0.001), and there was a higher proportion of ICU days per patient without sedation (RASS +1, 0, or -1; 50% vs 20%; P < 0.001). In addition, on routine assessment, patients were less often comatose (RASS -4 or -5) in the post-intervention period: 23% vs 65%, P < 0.001. However, median days per patient in which delirium was present increased (38% vs 20%, P < 0.01) during the post-intervention period. Average daily pain score per patient during the second period was very low at 0.5.

The above results document that the unit-level intervention to decrease sedation in patients with ALI was effective. However, the authors’ primary thrust in this article is to describe how they successfully implemented this cultural change in their ICU and successfully increased staff awareness and recognition of delirium. They point out that the observed increase in days of delirium per patient is unsurprising in the context of markedly reduced sedation administration and the prevalence of profound sedation, given that patients typically pass through a stage of delirium as they recover from sedative-induced coma toward normal wakefulness.

To implement their QI project, the authors first developed a new sedation protocol that targeted a RASS score of 0 ("alert and calm," see Table 1) in all ICU patients, including those on mechanical ventilation. The default route of administration for both benzodiazepines (e.g., midazolam) and narcotics (e.g., fentanyl) was to be intermittent dosing, and continuous infusions were to be permitted only if as-needed intermittent administration was inadequate according to specified criteria. Regardless of the route and schedule of administration, all sedatives and narcotics were to be discontinued daily at 8 a.m. Nurses were trained to anticipate agitated delirium at this time as patients awoke, and to tolerate a certain delirium level (RASS +2) for up to 4 hours before resuming intermittent sedation.

It is not surprising that the nurses were initially reluctant to comply with these elements of the protocol. However, Hager and colleagues went to great lengths to achieve buy-in by all staff members, with training in delirium recognition and treatment with both pharmacologic and non-pharmacologic approaches, and this was ultimately successful. The paper describes the QI process in detail. Briefly, the authors used the "4Es" model of Pronovost et al:2 engage, educate, execute, and evaluate. Engagement was achieved by an interdisciplinary QI team that included intensivists, MICU nurse educators, and pharmacists, with assistance from psychiatrists and a neuropsychologist. The day and night nursing staffs as well as staff and housestaff physicians received extensive training, and "super-users" who were available at all shifts were trained and directly facilitated the implementation of the protocol. The protocol was included in the electronic medical record, and bedside nurses reported each patient’s RASS score and delirium status (using the Confusion Assessment Method for ICU,3 CAM-ICU, see Table 2) during morning rounds. Evaluation of protocol acceptance and adherence by all ICU clinician groups continued on an ongoing basis following its formal implementation. No doubt as a result of this substantial and continuing effort by the interdisciplinary QI team, compliance with the protocol was very good, with at least 90% of scheduled RASS and CAM-ICU assessments being completed in both the control and QI periods.


In your ICU, is the need for sedation assessed using "clinical judgment," "experience," and "patient need" rather than by structured, objective assessment such as the RASS? Are sedatives and narcotics administered to mechanically ventilated patients routinely or "as needed" rather than according to a standardized protocol? Do such patients receive continuous infusions of these drugs rather than intermittent bolus dosing? Are sedative drips and intermittent dosing interrupted at least once each day in all patients receiving them, in order to determine by objective criteria whether they can be discontinued or their dosages substantially reduced? Are all patients in the ICU routinely and specifically evaluated for the presence of delirium?

In past decades, these questions could be thought of as relating to individual clinician preference and/or local practice traditions. Since then, however, they have been subjected to a great deal of clinical research, the results of which are clear and consistent. Administration of sedatives and narcotics to critically ill patients by continuous intravenous infusion results in giving them more drugs, and they remain longer both on the ventilator and in the ICU. The use of standardized, objective assessments of the need for sedation leads to less drug being administered. Surviving patients who received less sedation in the ICU have fewer neuropsychological and other problems in the succeeding months and years. And delirium in the ICU is strongly related both to the amount of sedation received and to a number of important long-term adverse outcomes.

Several measures have been well documented to reduce sedative use in the ICU. These include the use of objective, goal-directed assessment of the need for sedation, such as the RASS; intermittent bolus administration rather than continuous infusion; daily infusion interruption (until the patient awakens, not just moves); and efforts to use analgesics for discomfort rather than sedatives. The problem is not lack of evidence, but how to change long-established routine, clinician attitudes, and other aspects of local ICU culture. This study by Hager et al shows that these things can be changed, but also illustrates how much effort, time, and resources this requires. The finding that patients experienced more days of delirium during the period in which sedation was reduced, despite an overall reduction in the combined endpoint of coma and delirium, suggests that further research is needed to clarify and fine-tune best management in this important area of ICU care.


1. Sessler CN, et al. The Richmond Agitation-Sedation Scale:
Validity and reliability in adult intensive care patients. Am J Respir
Crit Care Med 2002;166:1338-1344.

2. Pronovost PJ, et al. Translating into practice: A model for large
scale knowledge translation. BMJ 2008;337:a1714.

3. Ely EW, et al. Delirium in mechanically ventilated patients: Validity
and reliability of the confusion assessment method for the
intensive care unit (CAM-ICU). JAMA 2001;286:2703-2710.