Management of the agitated intensive care unit patientedmedia.emory.edu/GStaton/Management of the...May 23, 2001 · scales, daily awakening, and other strat-egies take on more of

Management of the agitated intensive care unit patient

INTRODUCTION: AGITATIONROUNDTABLE MEETINGOVERVIEW

Agitation: 1. Violent motion. 2. Strongor tumultuous emotion.Management of the agitated patient is fastbecoming an area of major break-throughs for critical care medicine. Toillustrate, Figure 1 shows the total num-ber of articles found on MEDLINE using acombination of search words related tosedation and critical care. This crude sur-vey demonstrates an exponential rise inactivity surrounding this topic and helpssupport the view that study of agitation inthe critically ill patient is of rapidly ex-panding importance. Moreover, manage-ment of the agitated patient has devel-oped into an economically powerfulsubject, both for pharmaceutical compa-nies and for caregivers interested in im-proving the efficient use of intensive careunit (ICU) resources. It is increasinglyapparent that outcomes are significantlyinfluenced by the manner in which agi-tation is managed.

The quantity of articles being pub-lished is only part of the picture. Investi-gations related to agitation in criticalcare are yielding a variety of intriguingobservations including post-traumaticstress disorder and post-ICU depression,diagnosis of delirium, objective monitor-ing technology, sleep pattern changes,process/management strategies to en-hance clinical and economic outcomes,scoring systems, tailorability of therapeu-tic approaches, and bronchodilatory, an-tioxidant, and immunosuppressive prop-erties of sedative agents.

Rather than simply discussing strate-gies for sedation, it is the deliberate in-tent of this continuing education pro-gram to focus on the specific topic ofagitation (in the ICU patient). It is note-worthy that, although it is one of themost common issues facing critical care

practitioners, agitation in the ICU has noclear and concise definition.

The simple definition stated at the be-ginning of this article is from Funk andWagnall’s 1982. This explanation of “agi-tation” has merit because it encompassesboth physical and emotional distress. Un-der this characterization, either the non-sedated paralyzed patient or the comatosepatient with patient-ventilator asyn-chrony can be considered agitated, eventhough the two may represent oppositeends of a spectrum.

Accurate diagnosis of the cause of ag-itation frequently requires a careful anal-ysis of the patient’s history and physicalexamination, review of laboratory andother diagnostic data, knowledge of theeffectiveness of concomitant therapies,collaboration among members of theteam and family, and a good deal of ex-perience. The cause of agitation is oftenmultifactorial (e.g., pain and confusion ordelirium and withdrawal), and even withsuccessful management it is difficult tobe certain about precipitating factors inany single case. Anecdotes from patientsand clinicians can serve as powerful toolsfor the critical care team’s armamentar-ium and help increase understandingfrom both sympathetic and empatheticperspectives.

Pharmacologic management strate-gies for agitation include both preventionand treatment. Prevention commonlyguides the hand of the critical care clini-cian when a patient is being stabilizedand drips are ordered for analgesia andsedation in anticipation of agitation.Fine-tuning the therapy using agitationscales, daily awakening, and other strat-egies take on more of a treatment quality,as do pro re nata (PRN) agitation orders.Nonpharmacological approaches includea variety of environmental adjustmentsthat are frequently underutilized.

Yet, as obvious as these concepts fordefinition, diagnosis, and managementmay seem, it is difficult to consistentlyapply them to the literature (with theexception of short-term usage). There are

a number of well-designed and well-executed studies in longer-duration agi-tation management but, excluding thosein very focused populations (e.g., neuro-logic injury), most studies lump patientsinto groups for the purpose of assessingdiffering sedative regimens.

Comparative pharmaceutical trialshave been extraordinarily important toclinicians who deal regularly with agita-tion. These studies, as well as trials usinginnovative management techniques, arebecoming increasingly sophisticated inthe area of pharmacoeconomic assess-ment. There is still, however, a paucity ofcomprehensive studies evaluating the in-tegration of economic, clinical, and hu-manistic outcomes of agitated ICU pa-tients. Existing economic analysesinclude variables such as drug acquisitioncosts, ventilator duration, and ICU lengthof stay (LOS) to determine the “cost ef-fectiveness” of one drug regimen overanother; these are often only partial intheir scope. Assigning or assuming costsfor time in ICU or on a ventilator isfraught with the problems of evaluatingthe fixed and variable components. Op-portunity costs are usually ignored, asthey are exceedingly difficult to deter-mine. And, failure to include post-ICUcost and outcome information ignoresthe post-ICU morbidity that appearslinked to ICU sedation usage. These typesof problems with economic analyses arewidespread in the critically ill populationand are not unique to the topic of agita-tion management. Notwithstanding, itcan be said with a reasonable degree ofconfidence that the drug acquisition costof various regimens is only one—oftensmall—piece of the larger economic puz-zle.

Given the current tide of activity, it isconceivable that the approach to manag-ing agitation in the critically ill patientwill rise (or is rising) to a new level ofsophistication. At this new level, pharma-cologic and nonpharmacologic ap-proaches will be highly selective and fine-tuned to more precisely address theCopyright © 2002 by Lippincott Williams & Wilkins

S97Crit Care Med 2002 Vol. 30, No. 1 (Suppl.)

psychophysiologic disturbances found ineach critically ill individual. As a result,the critically ill population will experi-ence fewer side effects, shorter ICUcourses, better short-term and long-termoutcomes, improved cost-effectiveness ofcare, and reduced morbidity.

The fishbone/cause-and-effect dia-gram displayed in Figure 2 was designedto illustrate the challenge of managingagitation in the ICU patient by demon-strating schematically the interrelation-ship of many of the points presented inthis piece. Each item can be a significantfactor, and changing just one (e.g., uni-laterally starting a protocol) rarely worksunless careful thought is given to all theother variables. The cause-and-effect dia-gram is a quality improvement tool thatassists in identifying those variables.

The authors of this supplement are anexperienced, multidisciplinary group ofclinicians who discuss the topic of agita-tion from an academic and clinical per-spective spanning the development ofmodern critical care. The primary inten-tion of this continuing education pro-gram is to provide a practical frameworkfor managing agitation. It is hoped thatthe areas of controversy will be stimulat-ing to the reader. Two overriding ques-tions should emerge: What kind of evi-dence is needed to advance themanagement of agitation in the ICU?And, how do we bridge the evidence-caregap and put existing and emerging evi-dence into consistent daily practice?

Program Background

In June 2001, the authors participatedin a tele-roundtable meeting. Each author

presented his or her topic in depth followedby a brief question-and-answer period. Atthe end of the formal presentations, a seriesof questions was presented and the round-table discussion ensued. The questions,which were prepared in advance but notmade available to the authors until thetime of the conference, were meant to cre-ate controversy and offer brainstormingideas outside the structure demanded byscientific writing.

While finalizing their first drafts, theauthors were provided a draft of the revisedClinical Practice Guidelines for the sus-tained use of sedatives and analgesics,which is also published in this month’sissue of Critical Care Medicine. It is impor-tant to stress that this supplement is notmeant to supplant any of the recommenda-tions laid out by the Guideline Develop-ment Task Force. Opinions expressed inthis program may differ when the gradeof evidence is lower, however; thisgroup made no attempt to identify anevidence-based grade. Moreover, thiscontinuing medical education activitydoes not limit its scope with respect toduration of sedative use. Finally, thiswork is meant to be more speculative inits span.

Selectively, a number of topics werenot discussed in detail; these includeshock and sepsis. Although managing ag-itation in sepsis and shock is an essentialpart of care, it was felt that only general-ities could be addressed using the funda-mentals provided in the following sec-tions. Notwithstanding, observations ofdiffering effects of sedating agents on freeradicals and the immune system, for ex-ample, might lead to interesting break-

throughs in agitation management forcritically ill patients.

PHYSIOLOGY,PATHOPHYSIOLOGY, ANDDIFFERENTIAL DIAGNOSIS OFICU AGITATION

Agitation frequently occurs in criti-cally ill adult patients in the ICU and isassociated with potentially dangerouscomplications such as self-extubation, re-moval of arterial and venous catheters,increased systemic and myocardial oxy-gen consumption, and failure to partici-pate in therapeutic interventions (1, 2).The agitation syndrome may be caused bymany factors, including the underlyingillness itself, discomfort associated withinvasive catheters and tubes, and themany stimuli common to the ICU envi-ronment. Agitation develops regardless ofage, sex, or underlying diseases. The syn-drome complicates management in theICU, often leading to further morbidityand complications.

Definition, Symptoms, and Signsof Agitation

Although a simple definition of agita-tion in the critically ill patient is difficultto find, agitation can be described in sev-eral ways. Agitated patients exhibit con-tinual movement, characterized by con-stant fidgeting, moving from side to side,pulling at dressings and bed sheets, andattempting to remove catheters or othertubes. The agitated patient remains dis-oriented in one of several spheres. Theremay be a total lack of awareness as toname, place, or time. Alternatively, pa-tients may know who they are, but haveno idea of their current location. Depend-ing on the degree of agitation and theability of the patient to listen or commu-nicate, commands may or may not besuccessfully followed (4). The more com-plicated the request, the less likely thepatient will be able to respond in an ap-propriate manner. Patients capable ofcommunicating may exhibit intermit-tent, irrational thoughts or sentences.Within a long string of rambling conver-sation, some statements may make sensebut the vast majority of the conversationremains unintelligible (5). Shouting, call-ing out, or moaning can add to the clin-ical presentation. The agitated patientwill often exaggerate complaints of pain,when, in actuality, other factors such as

Figure 1. Number of articles on sedation in the intensive care unit.

S98 Crit Care Med 2002 Vol. 30, No. 1 (Suppl.)

the need to urinate or have a bowel move-ment are the causes of the complaints.

It is important to note that none of theabove descriptions characterize a patientundergoing neuromuscular blockadewho is agitated because of lack of seda-tion and analgesia. This condition, whichoften results in patients having vivid re-call while under pharmacologic paralysis,is a particularly disturbing occurrence tocritical care unit personnel, and mayhave long-term negative effects on thepatient.

Vital signs are generally abnormal in theagitated patient. Blood pressure may in-crease to dangerously high ranges, respira-tory rate may be elevated, and heart ratemay increase, with potential for ischemia(6). An elevated metabolic rate results in anincrease in overall oxygen requirementsand, if left to continue for a protractedperiod of time, an increase in caloric de-

mand. The agitated patient with a rapidrespiratory rate may not be able to synchro-nize respirations with the mechanical ven-tilator, resulting in high airway pressures,inadequate ventilation, and decreases inPO2 with either increases or decreases inPCO2, all of which further propagate thetendency toward agitation. These physio-logic changes frequently vary over 24 hrsdepending on the chronicity or intermit-tency of the agitation. Agitated patientsgenerally cannot concentrate or pay atten-tion to the caregivers around them, makingthe ability to follow requests or demandsexceedingly difficult.

Etiological Factors Contributingto Agitation

In the postoperative patient, the mul-tiple pharmacologic agents typically ad-ministered during the perioperative stage

can result in significant and often unpre-dictable interactions, leading to agitationand confusion. These agents include ben-zodiazepines, opioids, inhalation agents,anticholinergics, antibiotics, and musclerelaxants; they can interact in unpredict-able ways and may lead to a difficult man-agement situation, especially in the el-derly. In addition to drug– druginteractions, some agents alone, includ-ing lorazepam and anticholinergics, havebeen associated with the development ofagitation; once again, the aged are partic-ularly at risk (7). Frequently, the effectsof these drugs may not be related to theagent itself, but rather to multiple me-tabolites that have varying times of deg-radation and excretion (see Table 1).

A significant factor in the develop-ment of agitation in critically ill patients,predominantly in the postoperative pe-riod, is failure to provide adequate pain

Figure 2. Fishbone diagram of factors that may impact agitation.


control (8). In the United States, inade-quate pain management is often a resultof opioids being dosed at suboptimal lev-els because of concerns of respiratory de-pression and/or the development of de-pendence (9). However, these side effectsare unlikely over the short term if themedication is properly titrated to patientcomfort. Consequently, as clinicians wemust ensure that patients receive the ap-propriate dose necessary to achieve con-tinual pain relief.

Hypoxemia has long been associatedwith agitation. ICUs in most hospitalshave documented numerous clinical in-cidences in which hypoxemia had beenmisdiagnosed as agitation. PO2 levels of60 mm Hg or less (or oxygen saturationsbelow 90%) can contribute to agitationsecondary to hypoxemia. Hypotensionhas also been associated with agitationand is considered a form of brain injuryresulting from hypoperfusion. Likewise,hyper- and especially hypoglycemia canpromote severe agitation. Uremia and thepresence of elevated levels of heavy met-als such as lead, mercury, and manganesealso have been identified as causes of sig-nificant agitation in the critically ill pa-tient (4, 7).

Another cause of minor to severe agi-tation is brain injury, including closedhead trauma and bleeds from a ruptured

aneurysm with resulting subarachnoidhemorrhage. Thrombotic stroke maycause agitation as well. Brain abscesses,seizures, infections such as meningitis,and air embolism have all been associatedwith persistent and severe degrees of ag-itation (4, 7). A common situation in-volves frontal lobe injury following braintrauma, in which patients usually displayincreasing agitation, particularly as theybegin to awaken. Although difficult tocontrol, this increase in agitation canparadoxically be taken as a positive signin the patient’s recovery. Withdrawalfrom alcohol or from other agents includ-ing cocaine, opioids, and sedatives suchas benzodiazepines all contribute to braininjury and agitation (10). Cigarette smok-ers can suffer agitation from a lack ofnicotine. In many circumstances, with-out an adequate patient history, it may bedifficult to ascribe a cause for agitation.

Agitation can occur in patients whodevelop significant ventilator desynchro-nization. This is frequently caused by apoorly performing ventilator, with a delayin responding to the patient’s efforts atspontaneous breathing. Patients who re-quire short- or long-term intubation mayalso develop agitation, because of thestimulus of the endotracheal tube itself.Some intubated patients who are rela-tively alert become frustrated by their

inability to communicate and then evolveinto a cycle of continued agitation. Pa-tients frequently become anxious andtherefore agitated over the seriousness ofbeing critically ill. Finally, the ICU itself,with its high noise levels, lights, and con-tinual other stimuli, can significantlycontribute to further agitation (7, 11).

Differential Diagnosis ofAgitation

Agitated patients require that the cli-nician undertake a detailed work-up tofind and eliminate the various possiblecauses. At the top of any list, because ofits accompanying danger, should be hy-poxemia, which can be readily detectedby both arterial blood gas analysis andmeasurement of oxygen saturation (12).Of importance are occasions when a pa-tient with a low cardiac output state hasperfusion that is too low to maintain ad-equate oxygenation, resulting in hypox-emia resulting from cardiac dysfunctionrather than pulmonary dysfunction.

Metabolic abnormalities can usuallybe detected by laboratory analysis, in-cluding a basic electrolyte panel and de-termination of specific factors includingphosphate, calcium, and glucose levels. Itis often necessary to order additionaltests, not routinely performed, such as athyroid panel and liver function studies.Deficiencies in vitamin B-12, niacin, andthiamine should be considered, as well asheavy metal intoxication with lead, mer-cury, or manganese (4). A combination ofmedical history, physical findings, andappropriate laboratory testing will usu-ally identify a metabolic aberration.

Neurologic abnormalities often re-quire not only a detailed examination butalso a computed tomography (CT) scanand, in some cases, a magnetic resonanceimaging (MRI) scan. Undetected blood inthe brain after a bleed from an aneurysmor hypertensive bleed can cause signifi-cant agitation and an inability to respondappropriately to stimuli. An electroen-cephalogram (EEG) study may be usefulin the determination of diffuse encepha-lopathy, but is rarely specifically diagnos-tic. Nevertheless, in instances where pa-tients have received significant amountsof neurodepressant drugs such as barbi-turates or benzodiazepines, the EEG maybe a valuable diagnostic tool. Drasticallyelevated blood pressure in an agitated pa-tient should alert clinicians to a suspicionof hypertensive encephalopathy, a condi-tion requiring immediate control of

Table 1. Medications associated with agitation in patients in the intensive care unit (8)

Antibiotics Cardiac Drugs

Acyclovir CaptoprilAmphotericin B ClonidineCephalosporins DigoxinCiprofloxacin DopamineImipenen—cilastatin LabetalolKetoconazole LidocaineMetronidazole NifedipinePenicillin NitroprussideRifampin ProcainamideTrimethoprim—sulfamethoxazole Propranolol

Quinidine sulfate

Anticonvulsants Corticosteroids

Phenobarbital DexamethasonePhenytoin Methylprednisolone

Miscellaneous Drugs Narcotic Analgesics

Hydroxyzine CodeineKetamine MeperidineMetoclopramide Morphine sulfateTheophyllineAnticholinergicsBenzodiazepinesNonsteroidal anti-inflammatory agents


blood pressure as well as follow-up neu-rologic and CT examinations to rule outan intracranial bleed.

Obvious sources of pain, such as oper-ative procedures, are important causativefactors in the development of agitation.However, other less overt causes shouldnot be overlooked—pain from chesttubes, bladder spasm (which can developfrom the placement of catheters for uri-nary drainage), or injuries that may haveoccurred at the time of trauma.

Patients with chronic pain syndromes,such as low back pain, may become quiteuncomfortable when confined to one po-sition in a hospital bed in the ICU. Ob-taining an adequate history will assist inmaking this specific diagnosis.

Consequences of numerous drug in-terventions—drug reactions, drug inter-actions, and drug withdrawal—increasethe incidence of agitation in the ICU (6).The occurrence of undesirable drug–drug interactions should always be con-sidered when multiple drugs are beingused for pain, anxiety, and other psycho-biological issues. To diagnose an adversedrug interaction, it is often necessary tosequentially eliminate one or moreagents, or in some cases all agents. Eventhen, it may take several days for thedrugs and their metabolites to clear thepatient’s system before a positive re-sponse can be seen.

Infections can lead to agitation, butare more likely to manifest as increasedlethargy, with the patient becoming lessresponsive to stimuli and commands.One possible cause of infection in the ICUis direct bacterial or viral contaminationof the cerebrospinal fluid. Endotoxin re-lease from an ongoing illness may di-rectly affect brain function. It has beendemonstrated in patients with sepsis thatamino acid levels are commonly alteredboth in plasma and cerebrospinal fluid.Furthermore, normal brain metabolismcan be impaired in septic patients (7).Because sepsis is frequently associatedwith significant vasodilatation caused bythe release of nitric oxide, altered cere-bral perfusion may be an importantmechanism for abnormal brain metabo-lism. This problem must be viewed seri-ously, inasmuch as patients who developseptic encephalopathy appear to havetwice the mortality rate of other patients.

Renal and hepatic failure may alsolead to various levels of agitation andeven somnolence. Diagnostic features ofhepatic failure include neurologic dys-function with signs of encephalopathy

and triphasic waveforms seen on theEEG. Similar EEG changes can also bepresent in renal failure; however they arenot necessarily specific. Although pa-tients with hepatic and renal failure maybe agitated, they are usually not appro-priately responsive (12). Furthermore,control of agitation in these patientsmust be dealt with carefully because ofaltered metabolism and elimination ofpharmaceutical agents.

Last to be mentioned in the enumer-ation of differential diagnosis of agitationis nonclinical seizure activity, which maylead to significant degrees of agitationand may be difficult to differentiate fromtypical seizures. Usually, an EEG will benecessary to make the diagnosis. In pa-tients who have suffered an anoxic injury,a clonic seizure-like activity must also bedifferentiated from that which is second-ary to hypoxic injury to the brain, and notactually a seizure.

MONITORING AGITATION ANDBEDSIDE DECISION MAKING

Anxiety and agitation are common inthe ICU. Despite the frequency of theiroccurrence in the acutely ill patient, aclear definition, assessment strategy, ortreatment plan often remain unclear tothe bedside practitioner. Agitation is sub-ject to interpretation by the individualclinician, thereby making it difficult toobjectify and monitor from caregiver tocaregiver. Despite the proliferation of lit-erature in recent years, confusion stillexists among physicians, nurses, andother ICU staff with regard to a commondefinition of agitation, its incidence andcauses, the role of environmental factors,the relationship to ICU LOS, and the roleof drugs and interventions being em-ployed in the ICU. Establishing a multi-disciplinary standard of care for assess-ing, treating, and monitoring agitation inthe ICU is imperative for optimal patientmanagement and improved outcomes.

Anxiety/Agitation Continuum

In the critically ill patient, agitationcan be described along a continuum ofcontinuously changing physiologic stateswith varying behaviors and responses, af-fecting each patient differently within theseverity and complexity of their condi-tion. For most ICU practitioners, a verybrief description or assessment by an ex-perienced staff member at the bedsidecan provide a wealth of information for

reaching decisions with respect to thestatus of the patient, variables causingagitation, and intervention. The signs andsymptoms of agitation are fairly obvious.Descriptive terms commonly used in-clude restlessness; thrashing around inbed; pulling at catheters, tubes, and re-straints; overbreathing the ventilator;and asynchrony with the current ventila-tor settings. Abnormalities in vital signsinclude tachycardia, tachypnea, and hy-pertension.

Scales and Tools to MonitorAgitation

Patients in the ICU typically demon-strate complex disease states with a rap-idly changing hemodynamic status, mak-ing their requirements for treatment ofagitation fluctuate over time. These con-stantly changing requirements foster theneed for bedside clinicians to reassess andredefine the goals of therapy frequently.The ideal scale or tool to monitor agita-tion in the ICU should therefore be sim-ple to apply, yet describe clear gradedchanges between levels to allow titrationof interventions depending on the condi-tion of the patient.

Numerous scales and tools to monitorthe degree of agitation in clinical practiceare described in the literature. Most ofthese instruments attempt to evaluate asingle item, such as level of conscious-ness, at a single point in time. Otherscombine level of consciousness with de-scriptive responses to interventions, suchas mechanical ventilation. Unfortunately,there is no gold-standard method to eval-uate ICU patient response to agitationtherapy (13). Despite the weaknesses ofsome of the monitoring tools, applyingthem to protocol-driven interventionplans has been shown to improve patientoutcomes, such as duration of mechani-cal ventilation and ICU LOS (14).

The most commonly used scale in cur-rent literature is the Ramsay SedationScale (15). The Ramsay scale identifies sixlevels of sedation ranging from frank ag-itation to deep coma (see Table 2). De-spite its frequent use in research, theRamsay scale exhibits shortcomingswhen applied at the bedside of patientswith complex problems. The six levels ofsedation in the Ramsay scale are not mu-tually exclusive of one another; for exam-ple, the patient may appear to be asleepwith a sluggish response to glabellar tap(Ramsay 5) yet restless and anxious(Ramsay 1).


The Riker Sedation-Agitation Scale(SAS) was the first scale formally tested forreliability and validity in the ICU (see Table3). The SAS identifies seven symmetricallevels, ranging from dangerous agitation todeep sedation. This scale provides descrip-tions of patient behavior in varying levelsthat assist the bedside practitioner in dis-tinguishing between the levels (2).

The Motor Activity Assessment Scale(MAAS), which is similar in structure tothe SAS, uses patient behaviors to de-scribe the different levels of agitation(16). The MAAS identifies seven levelsranging from unresponsive to danger-ously agitated (see Table 4).

The Confusion Assessment Method forICU (CAM-ICU) described recently by Elyand colleagues (17) is being validated incritically ill patients with delirium (see Ta-ble 5). This tool for delirium has beentested in combination with a sedation scaleor the Glasgow Coma Scale (GCS) for chal-lenging patients and found to be simple toapply at the bedside, with inter-rater reli-ability, sensitivity, and specificity. The ef-fect on therapeutic intervention using thisscale is still being evaluated.

The development of noninvasive, ob-jective monitors of brain function usingEEG signals may lead to a more standard-ized assessment of agitation and sedation.This objective monitor is especially help-ful in the deeply sedated patient receivingneuromuscular blockade, as subjectivescales requiring patient input are notvalid. The Bispectral Index (BIS) providesa discrete value from 100 (completelyawake state) to �60 (deep sedation) and�40 (deep hypnotic state or barbituratecoma) by incorporating several EEGcomponents (18). Although the tech-nique has been shown to be a valid andreliable measure in the operating room(19), it has not been studied to any greatextent in the ICU. In one study designedto determine whether BIS correlates withresponses to commands during sedationand hypnosis induced by propofol, and tocompare BIS with targeted and measuredconcentrations of propofol in predictingparticipants’ responses to commands, 20volunteers were given propofol infusionsand EEGs were recorded for off-line anal-ysis of BIS. The results showed that theBIS is an accurate predictor of responseto verbal commands during sedation andhypnosis with propofol. Accuracy wasmaintained when propofol concentra-tions were increased or decreased andwhen repeated measurements were madeover time (20). Additional studies are

needed for BIS or other objective moni-toring tools before acceptance into clini-cal practice (18).

Bedside Decision-Making

Various factors and processes may influ-ence assessment and treatment practices inthe ICU. There is little in published litera-

ture concerning this topic. Weinart et al.(21) conducted focus group interviews withICU nurses at two hospitals and describedfactors affecting nurses’ delivery of sedativetherapy. Key factors identified as impactingsedative therapy included nursing attitudesand beliefs about critical illness, familymembers’ perception of agitation, andnurses’ workload and staffing ratios.

Table 2. Ramsay scale for assessing levelof sedation

Level Response

1 Patient awake and anxious, agitated, and/or restless2 Patient awake, cooperative, accepting ventilation, oriented, and tranquil3 Patient awake, responds to commands only4 Patient asleep; brisk response to light glabellar tap or loud auditory stimulus5 Patient asleep; sluggish response to light glabellar tap or loud auditory stimulus but does

respond to painful stimulus6 Patient asleep, no response to light glabellar tap or loud auditory stimulus

Table 3. Sedation-Agitation scale (2)

Score Diagnosis Description

7 Dangerous agitation Pulling at endotracheal tube, trying to remove catheters,climbing over bed rail, striking at staff, thrashing side toside

6 Very agitated Does not calm, despite frequent verbal reminding of limits,requires physical restraints, bites endotracheal tube

5 Agitated Anxious or mildly agitated, attempting to sit up, calmsdown to verbal instructions

4 Calm and cooperative Calm, awakens easily, follows commands3 Sedated Difficult to arouse, awakens to verbal stimuli or gentle

shaking but drifts off again, follows simple commands2 Very sedated Arouses to physical stimuli but does not communicate or

follow commands, may move spontaneously1 Unarousable Minimal or no response to noxious stimuli, does not

communicate or follow commands

Table 4. Motor Activity-Assessment scale (16)

Score Description Definition

0 Unresponsive Does not move with noxious stimulus1 Responsive only to noxious

stimuliOpens eyes or raises eyebrows or turns head toward

stimulus or moves limbs with noxious stimulus2 Responsive to touch or name Opens eyes or raises eyebrows or turns head toward

stimulus or moves limbs when touched or name isloudly spoken

3 Calm and cooperative No external stimulus is required to elicit movement,and the patient is adjusting sheets or clothespurposefully and follows commands

4 Restless and cooperative No external stimulus is required to elicit movement,and the patient is picking at sheets or clothes oruncovering self and follows commands

5 Agitated No external stimulus is required to elicit movementand attempting to sit up or moves limbs out of bedand does not consistently follow commands

6 Dangerously agitated,uncooperative

No external stimulus is required to elicit movement,and patient is pulling at tubes or catheters orthrashing side to side or striking at staff or tryingto climb out of bed and does not calm down whenasked


Teaching bedside staff the critical de-cision-making skills necessary to opti-mally manage agitation is an importantresponsibility of all critical care educa-tors. With respect to assessing the agi-tated patient, there are some simple con-siderations that need to be made quicklyand effortlessly by every bedside care-giver. One of the first things to be con-sidered is whether there is an underlyingphysiologic cause for the observed agita-tion symptoms. For example, disease-related pain and hypoxemia are two com-mon causes of agitation in the ICU.Interventions focused to correct the med-ical condition will therefore resolve theagitation. Another factor to be consideredon initial assessment is the possibility ofany ongoing therapy being the cause ofthe agitation. For example, the patientmay be exhibiting a medication-relatedside effect, a malfunctioning nasogastrictube causing feelings of nausea and agi-tation, or a blocked Foley catheter. Otherinitial considerations must include thepossibility that agitation may be a resultof withdrawal symptoms from eithermedications administered before ICU ad-

mission or abuse of alcohol or illicitdrugs. After exclusion of obvious causesof agitation, considerations with regardto the hemodynamic stability of the pa-tient will affect speed of bedside staff in-tervention, and the determination of re-quirements for immediate pharmacologictherapy or, alternatively, whether non-pharmacologic strategies may be appro-priate to treat agitation (8, 11, 22).

Once the bedside staff has ruled outobvious causes and identified the severityof the agitation, considerations regardingoptimal interventions can be made to en-sure the best patient outcomes. Pharma-cologic agents such as benzodiazepinesor propofol are frequently administeredin the ICU to treat agitation; however,most bedside caregivers also employ non-pharmacologic interventions. These in-terventions include optimizing commu-nication with the patient, coaching thepatient in relaxation techniques, reori-enting the patient to the unit, reducingenvironmental stimuli and noise, andproviding psychosocial support (6). Crit-ically ill patients exhibit severe sleep frag-mentation and reduced restorative sleep

with suppression of rapid eye move-ment. The exact etiology and patho-physiology of sleep disruption in theICU remains unknown. Regardless ofthe cause, serious adverse effects are as-sociated with sleep deprivation, includingimpaired immunity, impaired protein syn-thesis, respiratory abnormalities, and dis-rupted thermoregulation. Patients in theICU often consider sleep disruption to beone of the most unpleasant aspects of theirillness (23).

Patient-specific goals for therapy canbe defined to ensure desired endpoints.These goals are often linked to the in-dications for therapy—for example,treatment of anxiety or agitation, abol-ishing discordance with the ventilator,reducing oxygen consumption, or as anadjunct to neuromuscular blockingagents.

Establishing and ImplementingSedation Guidelines andProtocols

The successful development and im-plementation of sedation guidelines and

Table 5. Confusion assessment method for the intensive care unit (17)

Feature 1. Acute onset of mental status changes or fluctuating courseIs there evidence of an acute change in mental status from the baseline?Did the (abnormal) behavior fluctuate during the past 24 hrs, that is, tend to come and go or increase and decrease in severity?Did the sedation scale (e.g., Sedation-Agitation scale or Motor Activity-Assessment scale) or coma scale (Glasgow Coma scale) fluctuate in the past

24 hrs?Feature 2. Inattention

Did the patient have difficulty focusing attention?Is there a reduced ability to maintain and shift attention?How does the patient score on the Attention Screening Examination, or ASE (i.e., visual component ASE tests the patient’s ability to pay attention

via recall of ten pictures; auditory component tests attention via having patient squeeze hands or nod whenever the letter “A” is called in arandom letter sequence)?

Feature 3. Disorganized thinkingIf the patient is already extubated from the ventilator, determine whether the patient’s thinking is disorganized or incoherent, such as rambling or

irrelevant conversation, unclear or illogical flow of ideas, or unpredictable switching from subject to subject.For those still on the ventilator, can the patient answer the following four questions correctly?

Will a stone float on water?Are there fish in the sea?Does 1 pound weigh more than 2 pounds?Can you use a hammer to pound a nail?

Was the patient able to follow questions and commands throughout the assessment?“Are you having any unclear thinking?”“Hold up this many fingers” (examiner holds two fingers in front of patient).“Now do the same thing with the other hand” (not repeating the number of fingers).

Feature 4. Altered level of consciousnessAny level of consciousness other than alert (e.g., vigilant, lethargic, stupor, or coma).

Alert: Normal, spontaneously fully aware of environment, interacts appropriatelyVigilant: HyperalertLethargic: Drowsy but easily aroused, unaware of some elements in the environment, or not spontaneously interacting appropriately with

the interviewer; becomes fully aware and appropriately interactive when prodded minimallyStupor: Difficult to arouse, unaware of some or all elements in the environment, or not spontaneously interacting with the interviewer;

becomes incompletely aware and inappropriately interactive when prodded strongly; can be aroused only by vigorous andrepeated stimuli and as soon as the stimulus ceases, stuporous subject lapses back into the unresponsive state

Coma: Unarousable, unaware of all elements in the environment, with no spontaneous interaction or awareness of the interviewer, sothat the interview is impossible even with maximal prodding

Patients are diagnosed with delirium if they have Features 1 and 2 as well as either Features 3 or 4.


protocols require multidisciplinary inputand additional training for all caregivers;physicians and nurses need to agree onmonitoring scales and tools and then in-sure that these scales are used reliablyacross disciplines and within units. It isessential to determine specific details re-garding the frequency of assessment, pre-defined end points of therapy, and evalu-ation of patient outcomes. Forms andflow sheets currently in use at the bedsidecan be used for developing documenta-tion systems. Using these documentationsystems to foster communication be-tween disciplines (e.g., nurse to physi-cian) and within disciplines (shift to shift)assures uniformity of guidelines. Devel-opment of drug administration guide-lines that foster current pharmacologic/pharmacokinetic recommendations andstandards for acutely ill patients is en-couraged.

Brook et al. (14) conducted a ran-domized, controlled trial of patients ina medical ICU that compared protocol-directed with nonprotocol-directed se-dation administration. Patients in theprotocol-directed group had less timeon mechanical ventilation, shorter ICULOS, and shorter hospital LOS, as wellas decreased need for tracheostomycompared with those in the nonproto-col-directed group. These results dem-onstrate that using a multidisciplinary-designed sedation protocol can improvepatient outcomes and decrease overallcost. Other bedside strategies to opti-mize outcomes in patients receivingtherapy for agitation in the ICU includeinstituting daily reassessment and in-terruptions of sedative infusions (23).Daily interruption of sedative infusionswas found to decrease duration of me-chanical ventilation (4.9 days comparedwith 7.3 days), decrease ICU LOS, andimprove clinicians’ ability to performdaily neurologic examinations, there-fore reducing the need for diagnosticstudies to evaluate unexplained alter-ations in mental status.

Regulatory Issues

The Joint Commission on Accredita-tion of Healthcare Organizations(JCAHO) reinforces the importance of ap-propriate sedation in its revised Stan-dards and Intents for Sedation and Anes-thesia Care, effective January 1, 2001(24). Institutional compliance with theserevised standards requires the institution

to ensure that all individuals administer-ing sedation be qualified and have appro-priate credentials to manage patients re-ceiving moderate or deep sedation. Inthese revised standards, the levels of se-dation have been defined by JCAHO asfollows:

● Minimal sedation is a drug-inducedstate during which patients respondnormally to verbal commands, al-though cognitive function and coordi-nation may be impaired; ventilatoryand cardiovascular functions are unaf-fected.

● Moderate sedation is a drug-induceddepression of consciousness duringwhich patients respond purposefully toverbal commands, either alone or ac-companied by light tactile stimulation.No interventions are required to main-tain a patent airway, and spontaneousventilation is adequate. Cardiovascularfunction is usually maintained.

● Deep sedation is a drug-induced de-pression of consciousness duringwhich patients cannot be easilyaroused but respond purposefully fol-lowing repeated or painful stimulation.The ability to independently maintainventilatory function may be impaired.Patients may require assistance inmaintaining a patent airway, and spon-taneous ventilation may be inadequate.Cardiovascular function is usuallymaintained.

Verification of compliance with thestandard requires the institution toprovide monitoring standards and as-sessment tools within its policies forcare. Institution-wide agreement re-garding the standard of care for seda-tion practice requires ensuring thecompetency of all staff caring for pa-tients requiring sedation. Evidence ofmultidisciplinary teaching strategiesincluding sedation assessment parame-ters, documentation tools, and evalua-tion of patient outcomes is suggestedthroughout all areas of the institutionin which sedation is administered(CAMH update, 3 August 2000: Compre-hensive Accreditation Manual for Hos-pitals, effective 1/1/01). Application ofthe current critical care literature, in-cluding the use of protocols, algo-rithms, assessment tools, and deliverystrategies, reinforces these regulatorystandards.

ICU SEDATIVE PHARMACOLOGYUPDATE: A REVIEW OFCOMMONLY USED ANDEMERGING AGENTS

Analgesics and sedatives are mainstaysof supportive patient care in the ICU.Critically ill patients are frequently inpain as a result of their medical conditionor surgery; mechanical ventilation andenvironmental factors cause additionalstresses. Delirium and other adverse ef-fects of the ICU stay necessitate the use ofsedation to prevent or alleviate the agita-tion that commonly results. Analgesia isimportant for the same reason: severepain is a frequent cause of agitation anddelirium.

It is generally recommended that pa-tients in the ICU receive sufficient anal-gesia, usually with opiates, before seda-tives are administered. Traditionally,benzodiazepines such as midazolam,lorazepam, and diazepam have been usedfor sedation, whereas haloperidol hasbeen used to treat delirium. More re-cently, propofol has become a populardrug for ICU sedation; the introduction ofemerging sedative agents, such as dexme-detomidine and potentially 2% propofol,will further broaden clinician options.

The choice of an appropriate sedativeis often difficult, and depends on the in-dividual needs of the patient. For exam-ple, if rapid awakening to a state of alert-ness is required, as in the neurologicpatient who requires frequent monitor-ing, propofol is the preferred agent. Forlong-term sedation, lorazepam is consid-ered the drug of choice. Haloperidol isthe preferred agent for delirium. It isessential that practitioners become famil-iar with the properties and uses of theseagents so that the patient is given theopportunity for the best outcome.

Maintenance of adequate sedation is akey component of ICU care. Ventilatorysupport frequently induces anxiety, pain,and asynchrony. Appropriate sedativesand analgesics can alleviate much of thisdiscomfort, and can lessen stress-inducedincreases in oxygen consumption. In pa-tients with respiratory failure, the admin-istration of sedatives at appropriate doseshelps increase chest wall compliance, al-lows the manipulation of inspiratory toexpiratory ratio and other variables, im-proves oxygenation, and reduces desyn-chronized breathing (25, 26).

Alleviation of pain is an equally impor-tant component of care in the ICU. Anincreased level of pain activates the sym-


pathetic nervous system, placing addi-tional demands on the cardiovascular sys-tem in critically ill patients. When pain isprolonged, it contributes to severe anxi-ety and even delirium. The hypermeta-bolic state after injury is exacerbated bypain, potentially leading to diminishedimmune function and impaired woundhealing. Therefore, adequate analgesia isof essential importance in the manage-ment of these patients (27).

The primary goals of sedative therapy,once a pain-free state is achieved, areanxiolysis, hypnosis, and amnesia. Not allsedative agents used in the ICU canachieve these goals, making the correctchoice of a sedative of paramount impor-tance. Similar plasma concentrations of agiven sedative can have varied results indifferent individuals with respect to drugdisposition and pharmacodynamic effect.The doses of drug required for adequatesedation also change during the ICU staybased on the nature and course of thedisease, interaction of the sedative withother pharmacologic agents, and the re-sponse to therapy. No single depth ofsedation or single sedative agent is appro-priate for all patients (27).

Sedatives are not used only for seda-tion in the ICU; other indications includemanagement of drug withdrawal syn-dromes and treatment of seizures. Properuse of these agents can enhance patientcomfort and safety, but, if inappropriatelychosen or incorrectly administered, theoccurrence of side effects can lead to in-creased morbidity, mortality, and costs(28) (Table 6).

The practice parameters for intrave-nous (iv) sedation in the ICU published in1995 by the American College of CriticalCare Medicine (ACCM) and the Society ofCritical Care Medicine (SCCM) have been

updated to include an evaluation of theliterature published since 1994 compar-ing the use of sedatives and analgesics inthe ICU. Now known as Clinical PracticeGuidelines, the 2001 guidelines recom-mend that sedation of critically ill pa-tients be started only after provision ofadequate analgesia and treatment of re-versible physiologic causes. For rapid se-dation of acutely agitated patients, mida-zolam or diazepam should be used.Propofol is the preferred sedative whenrapid awakening (as for neurologic as-sessment or extubation) is important(29). Midazolam is recommended forshort-term use only, as it produces un-predictable awakening and/or time to ex-tubation when infusions continue formore than 48–72 hrs. For intermittent ivdoses or continuous infusion, the recom-mended drug for sedation in most pa-tients is lorazepam (30, 31). Haloperidolis the preferred agent for the treatment ofdelirium in critically ill patients (32).

Guidelines have been implemented tostandardize care and lower costs, and anincreasing number of hospitals haveadopted them for use in ICU sedation.Mascia et al. (33) and Devlin et al. (34)examined the impact of guidelines oncosts and outcomes.

Mascia et al. (33) performed a prospec-tive cost-effectiveness analysis. Trackingof 72 eligible baseline (preguidelines) pa-tients was followed by the developmentand introduction of guidelines developedwith multidisciplinary input, along withan academic detail process to promotetheir use. Several months following theintroduction of these guidelines, a secondgroup of 84 follow-up (postguidelines)patients was tracked. Both groups weresimilar with regard to number of regi-mens and days of treatment. Ventilator

time and LOS were shorter in the post-guidelines group, without a compromisein quality of care, and drug costs weresignificantly reduced in the postguide-lines group. The costs of propofol whengiven for 24 hrs or longer, for example,were $355.82 to $1,010.85 for thepreguidelines group and $123.06 to$460.50 for the postguidelines group. To-tal sedation costs were reduced from$4,515 to $1,152 (p � .081) (33).

The Devlin study (34) was designed asa before-and-after study in a 15-bed med-ical-surgical ICU. Guidelines were devel-oped through a consensus of physicians,nurses, and pharmacists. Fifty patientswere evaluated before the guidelines weredeveloped, and 50 were evaluated afterthe guidelines were implemented. Theguidelines promoted the use of loraz-epam over midazolam, with propofol sug-gested for patients not successfully se-dated with high-dose lorazepam,haloperidol, or morphine. Over the2-month study period, there was no dif-ference in the median weaning time forthe two groups. Total sedation costs,however, decreased from $4,515 in thepreguidelines group to $1,152 in thepostguidelines group (p � .081). The me-dian per-patient sedation drug cost de-creased from $11.27 (range, $0–1,340) inthe preguidelines group to $3.55 (range,$0–250) in the postguidelines group. Thenumber of postguidelines patients receiv-ing continuous infusions was signifi-cantly less than preguidelines patients(14% vs. 56%, respectively; p � .05). Al-though it did not reach significance,there was a trend for fewer postguidelinespatients to receive neuromuscular block-ing agents in the ICU (4% vs. 8%). Thisstudy demonstrated that high compliancewith ICU sedation guidelines led to a 75%decrease in sedation drug costs (34).

Opioids

Opioids are the primary agents usedfor analgesia in the ICU. They are lipid-soluble and bind to opiate receptors inthe central and peripheral nervous sys-tem. At low doses, opioids provide anal-gesia but not anxiolysis, whereas athigher doses they act as sedatives. Allopioids share therapeutic properties butvary in potency and pharmacokinetics.Morphine, but not fentanyl, induces hista-mine release, which results in hypotension.Although opioids can be given by severalroutes, the iv method is preferred in theICU for reliable drug delivery. When given

Table 6. Properties of an ideal sedative (29–31)

Easily titratable level of adequate sedationRapid onset of actionShort acting, allowing patient assessment, easy weaning from mechanical ventilation, and early

extubationNo adverse effectsNo nausea, vomiting, phlebitisNo anaphylaxis or allergic reactionMinimal metabolism; not dependent on normal hepatic, renal, or pulmonary functionNo active or toxic metabolitesNo suppression of cortisol production by the adrenal cortexNo interactions or incompatibilities with other commonly prescribed intensive care unit drugsEase of administrationLack of accumulation with prolonged administrationDoes not promote growth of pathogensCost effectiveEasily prepared and long shelf-life


in iv therapeutic doses, opioids cause seda-tion, in the sense of a clouded sensorium.They do not, however, possess amnesticproperties (9, 35, 36).

Opioids are stereospecific agonists at en-dorphin receptor sites in the central ner-vous system and other tissues. Mu-1 recep-tors are believed to mediate the supraspinalanalgesic action of opioids, whereas ago-nism at mu-2 receptor sites is thought toproduce side effects including ventilatorydepression, bradycardia, and physical addic-tion. All drugs in this class primarily un-dergo hepatic metabolism. Aside from an-algesia, an important neurophysiologiceffect of opioids is respiratory depression.The respiratory rate, minute ventilation,and the sensitivity of the medullary respi-ratory center to CO2 all decrease after ad-ministration of opioids (37).

Morphine sulfate is the prototypic opi-oid and is the preferred opioid analgesicin patients with stable hemodynamics. Ithas lower lipid solubility than does fen-tanyl; the result is a delayed onset ofaction. Morphine induces the release ofhistamine, which increases the likelihoodof hypotension secondary to vasodilata-tion (9, 39). A metabolite of morphine,morphine-6-glucuronide, is excreted inthe urine and may accumulate in renalfailure. The opiate activity of this metab-olite is several times greater than that ofmorphine, and its accumulation in pa-tients with renal failure has been re-ported to prolong narcosis (37).

Fentanyl citrate, a synthetic narcoticanalgesic up to 100 times more potentthan morphine, is highly lipid-solubleand has a rapid onset of action because itquickly crosses the blood-brain barrier.This drug has no active metabolites and isnot associated with histamine release orvenodilating effects. Because of thesecharacteristics, fentanyl is the recom-mended opioid as second-line therapy inpatients with unstable hemodynamics orthose who cannot tolerate the adverseeffects of morphine. Fentanyl should beadministered by continuous infusion for

sustained effect because of its short du-ration of action (35, 38).

Hydromorphone is a highly potentopioid with no active metabolites. Hydro-morphone can be used during shortagesof fentanyl because it has no active me-tabolites and does not cause clinically sig-nificant histamine release. Remifentanil,an extremely short-acting opioid analge-sic with a rapid onset of action, is rarelyused in the ICU setting. Meperidineshould be avoided in the ICU because ofthe neuroexcitatory properties of its me-tabolite that accumulates in renal failure(35, 39).

The use of opioids is associated withundesirable side effects. Because all opi-oids produce respiratory depression,weaning may be difficult in patients re-ceiving these agents. The incidence ofhypotension varies with the opioid and itsproperties with respect to vasodilatationand histamine release. Gastrointestinalside effects include slowing of gastroin-testinal motility; this can lead to ileus,gastric distention, nausea, and vomiting.Naloxone, an opioid antagonist, is muchshorter-acting than most opioids and isthe most widely used narcotic antagonistin the ICU for reversal of side effects.Dependence and withdrawal can be aproblem in patients receiving long-termopioid therapy in the ICU. Sudden dis-continuation of therapy to prepare a pa-tient for extubation may result in thedevelopment of withdrawal symptoms.Tapering the dose, while monitoring forsigns of withdrawal, is recommended inall ICU patients who have been on long-term opioid therapy (9, 35).

Benzodiazepines

The class of agents most widely usedfor sedation in the ICU is the benzodiaz-epines (see Table 7) (40). These drugsprovide anxiolysis and amnesia, but theyhave no analgesic properties. The twopredominant mechanisms of action ofbenzodiazepines within the nervous sys-

tem involve activity at �-aminobutyricacid (GABA) receptors. Potentiation ofGABA-mediated transmission by benzodi-azepines is apparently responsible for thesomnolent, anxiolytic, and anticonvul-sant actions, whereas the amnestic prop-erty seems to correlate with GABA ago-nist activity in the limbic cortex (38). Thebenzodiazepines currently used in theICU setting are diazepam, lorazepam, andmidazolam. The primary difference be-tween these agents relates to their phar-macokinetics.

The liver extensively clears benzodiaz-epines. The effects of these drugs may beprolonged in critically ill patients becauseof decreased metabolism or in the pres-ence of severe liver disease. Because ben-zodiazepines are sequestered in fat stores,prolonged sedation may occur withchronic administration (37). The effectsof benzodiazepines can be reversed byflumazenil, a competitive antagonist witha rapid onset and relatively short dura-tion of action in comparison with theprolonged effects of benzodiazepines(38).

Withdrawal syndromes are known tooccur after continued use of benzodiaz-epines, and tachyphylaxis can developwithin hours to days. The latter requireseither dose escalation or use of anothersedative agent. After several weeks of con-tinued use, the acute cessation of therapycan give rise to a syndrome that mani-fests as tremors, diaphoresis, photopho-bia, insomnia, abdominal discomfort, hy-pertension, and seizures (37).

Diazepam is a long-acting lipophilicbenzodiazepine that rapidly penetratesthe central nervous system, so that seda-tive effects are seen within 2–3 mins.Although diazepam is no longer recom-mended for routine use in the ICU, thereare reports of its use for long-term seda-tion in selected patients (32). This recom-mendation is the result of a scheduledintermittent dosing regimen that mayeasily lead to excessive and prolongedsedation. Also, dilution is needed for con-tinuous infusion, and this usually re-quires large volumes of fluid administra-tion. Other disadvantages of diazepam arethe common occurrence of pain andthrombophlebitis when the drug is ad-ministered by peripheral vein injection(39). Diazepam has an active metabolite,dimethyl-diazepam, which is only slightlyless potent than diazepam and has anelimination half-life of 96 hrs, longerthan that of the parent compound (37).

Table 7. Pharmacokinetics of diazepam, lorazepam, midazolam, and propofol in healthy volunteers

Diazepam Lorazepam Midazolam Propofol

Half-life (�), min 30–66 3–20 6–15 2–3Half-life (�), hrs 24–57 14 1.7–2.6a 0.5–1.0Volume of distribution, L/kg 0.7–1.7 1.14–1.3 1.1–1.7 5.4–7.8Clearance, mL/kg/min 0.24–0.53 1.05–1.1 6.4–11.1 26–29Protein binding, % 96–99 86–93 97 98Active metabolites Yes No Yes No

aUp to 30 hrs in patients in the intensive care unit. Adapted from Young, 2000, Table 3.


Lorazepam, an intermediate-actingbenzodiazepine, is less lipophilic than di-azepam and therefore has less potentialfor accumulation. The drug is usually ad-ministered by intermittent iv injection,but continuous infusion may be used.Because there is a slight delay in theonset of action of lorazepam, it is accept-able to administer a single dose of a morerapidly acting benzodiazepine whenachievement of rapid sedation is neces-sary. Compared with midazolam, loraz-epam is longer acting, causes less hypo-tension, produces equally effectiveanterograde amnesia, and, with pro-longed administration, produces morerapid awakening (39). The new ClinicalPractice Guidelines recommend loraz-epam for the sedation of most patients byintermittent iv doses or continuous infu-sion (32). The drug has no active metab-olites and its metabolism is less affectedby advanced age or liver dysfunctioncompared with midazolam. Lorazepam isassociated with a stable hemodynamicprofile, even when opioids are concur-rently administered (41). It may, how-ever, be unstable in solution and can pre-cipitate in iv catheters and tubing,particularly if infusions last longer than12 hrs. This can add to the cost of ther-apy. Propylene glycol toxicity, marked byacidosis and renal failure, has occurredwith higher doses of lorazepam or pro-longed infusion of the drug (27). Therewas recently a case report in Pharmaco-therapy of propylene glycol toxicity oflorazepam in only 3 days in a patient withrenal failure (first case in �72 hrs oftherapy (42).

Midazolam is a short-acting, water-soluble benzodiazepine that is trans-formed to a lipophilic compound in theblood. The drug rapidly penetrates thecentral nervous system to produce ashort onset of sedation of 2–5 mins. Itsduration of effect is brief because it israpidly redistributed, a property that fa-vors continuous infusion for mainte-nance of sedation (39). Use of midazolamfor chronic sedation is limited because, insome patients, there is prolonged elimi-nation half-life of up to 30 hrs and asso-ciated variability in the time of return toconsciousness after discontinuation;however, few adverse hemodynamic andrespiratory effects are seen with theshort-term use of midazolam. To mini-mize the incidence of withdrawal phe-nomena after long-term duration infu-sions, the drug should be properlytapered (41). The new Clinical Practice

Guidelines recommend midazolam forrapid sedation of acutely agitated pa-tients. It is recommended for short-termuse only, as it produces unpredictableawakening and/or time to extubationwhen infusions continue for more than48–72 hrs (32).

Midazolam exhibits dose-related hyp-notic, anxiolytic, amnestic, and anticon-vulsant actions. The drug also causesdose-related respiratory depression, andat large doses can cause hypotension andvasodilatation. When midazolam is ad-ministered as a continuous infusion,however, these effects are minimal (27).The drug is biotransformed to an activemetabolite in the liver that is not as po-tent and is shorter-lasting than the par-ent compound. Because only small quan-tities are formed during continuousinfusion of midazolam, this metabolitedoes not contribute significantly to thepharmacologic activity of the drug (ex-cept in patients with severe renal failure)(27). The metabolism of midazolam isreduced when administered to patientsreceiving cytochrome P-450 3A4 inhibi-tors such as erythromycin and flucon-azole (43).

Midazolam infusions for sedation havebeen compared with other benzodiaz-epines. In a prospective randomizedstudy, Pohlman et al. (44) compared theefficacy of continuous infusions of mida-zolam (mean dose, 0.24 mg/kg/hr) andlorazepam (mean dose, 0.06 mg/kg/hr)for sedation of mechanically ventilatedpatients in a medical ICU. For both drugs,the time to achieve sedation was oftenprolonged, and higher doses than thosereported in the literature were requiredto maintain sedation. Time to awakeningwas occasionally delayed for more than24 hrs after discontinuation of either in-fusion, and large volumes of fluid wereneeded to deliver the required doses. Inaddition, patients treated with midazo-lam had a tendency to return more slowlyto baseline mental status. There wasequally effective sedation and no differ-ence in other clinical variables. The rela-tive potency of lorazepam was two to fourtimes greater than that of midazolam.Despite a standard protocol for sedationin this study, the mean time to achieveadequate sedation was 115 mins for theentire study group, which the investiga-tors suggest was the result of patientdose-effect variability, poorly developeddosing guidelines, and the changing clin-ical condition of ICU patients.

Propofol

Propofol is a sedative-hypnotic withno analgesic action (37); it has sedative,hypnotic, and anxiolytic properties (39).Other effects of propofol are bronchodi-lation, seizure suppression, muscle relax-ation, and possible anti-inflammatoryand antiplatelet effects. Propofol is highlyfat soluble, and hence is formulated in anintralipid, a 1% emulsion containing10% soya bean oil, 2.25% glycerol, and1.2% purified egg phosphatide (37). A 2%formulation of propofol is currently un-der Food and Drug Administration (FDA)evaluation.

After a single iv dose, the onset ofaction of propofol is rapid (1–2 mins) andits effect is brief (10–15 mins) because ofrapid central nervous system penetrationand subsequent redistribution. Therefore,propofol is administered only by contin-uous infusion when used for sedation.Long-term infusion results in accumula-tion within lipid stores, so that there is aprolonged elimination phase with a half-life of up to 300–700 mins. However,subtherapeutic plasma concentrations ofthe drug are maintained after discontin-uation because of rapid clearance, thuslimiting the clinical significance of thishalf-life value (39). Although the mecha-nism of action of propofol is still notcompletely understood, the drug appearsto activate the GABA-A receptor withinthe central nervous system. Propofol al-ters the sensorium in a dose-dependentmanner, from light sedation to generalanesthesia. The drug is also a potent re-spiratory depressant, causing a reductionin systemic vascular resistance and pos-sibly hypotension, especially when ad-ministered as a bolus. Parallel with itsaction on the level of arousal, propofoldecreases cerebral metabolism, which re-sults in a coupled decline in cerebralblood flow and a decrease in intracranialpressure. Sedative infusion doses of thisagent typically result in minimal hemo-dynamic alteration with no change inperfusion pressure as long as adequateintravascular volume status is main-tained (37).

Propofol is considered an ultra short-acting agent for two reasons. Because it ishighly lipophilic, the drug redistributesto fatty tissues to such an extent that itsvolume of distribution approaches 600–800 L. Second, drug clearance is calcu-lated to be more than 1.5–2.0 L/min,exceeding hepatic blood flow and sug-gesting possible extrahepatic metabo-


lism. These kinetics result in a very rapiduptake and elimination from plasma withlittle accumulation and a low likelihoodof delayed recovery from sedation. De-spite maintenance of propofol sedationfor up to several days, recovery to anawake and responsive state after discon-tinuation of therapy occurs within 10–15mins (37). The pharmacokinetics ofpropofol are not altered in patients withrenal or hepatic disease (36).

The use of propofol is not currentlyrecommended for pediatric patients inthe ICU because of reports of metabolicacidosis with accompanying lipemic se-rum, bradyarrhythmias, and fatal myo-cardial failure; this occurred in patientsbeing treated with excessively high doses(45). In adults, prolonged high-dose infu-sion may also lead to cardiac failure (31,46).

Several studies have compared mida-zolam with propofol infusions for seda-tion in medical, surgical, and coronaryICUs (47–58). Both drugs are generallysafe and effective in the early postopera-tive period. Patients sedated with propo-fol infusions recover more rapidly, withless variability in recovery times, com-pared with patients sedated with midazo-lam infusions. Furthermore, alterationsin the level of sedation are controlledmore easily with propofol than with mi-dazolam infusions. There is no differencein the quality of sedation. In patientstreated with propofol, especially with aloading dose, there has been observed anincreased incidence of hypotension com-pared with midazolam, and therefore abolus is not recommended (41). In mostof these studies, the time from drug dis-continuation to successful ventilatorweaning was significantly shorter for pa-tients receiving propofol.

Barrientos-Vega et al. (59) conductedan open-label, randomized, prospective,phase IV clinical trial to evaluate the im-pact of prolonged sedation of critically illpatients with midazolam or propofol onweaning and ICU costs, using a cost-of-care approach. This trial, conducted inthe medical and surgical ICU of a com-munity hospital in Spain, included 108patients requiring mechanical ventilationfor at least 24 hrs. Although both drugsprovided equivalent sedation, administra-tion of propofol was associated with ashorter weaning time than midazolam,resulting in a more favorable economicprofile. The midazolam group showed ahigher rate of patients exhibiting inade-quate sedation, whereas the propofol

group showed a higher rate of therapeu-tic failure when cases of hypertriglyceri-demia were factored in. Neither differ-ence reached statistical significance.Propofol infusion was also associatedwith an earlier extubation time than mi-dazolam, including not only the time toawakening, but also time from the firstT-bridge trial to extubation.

The same investigators conducted an-other study comparing propofol 2% andpropofol 1% with respect to effectivenessand wake-up time required for prolongedsedation. Results were then comparedwith the results of the earlier study com-paring propofol 1% and midazolam. Se-dation with either propofol formulationwas associated with a more rapid weaningtime and more predictable wake-up thansedation with midazolam, although thedifferences did not reach statistical sig-nificance. The cost-effectiveness profile ofboth propofol concentrations was betterthan that of midazolam. Differences weresignificant for up to 288 hrs of sedationfor the propofol 1% group and up to 312hrs of sedation for the propofol 2%group. The economic benefits of propofolvs. midazolam were associated withshorter weaning time and shorter ICUstays, whereas the economic benefits ofpropofol 2% were associated with re-duced frequency of hypertriglyceridemiacompared with propofol 1% (60).

Carrasco et al. (61), in another trialconducted in Spain, compared the effi-cacy, safety, and cost of propofol and mi-dazolam for short-, medium-, and long-term sedation of critically ill patients.The study randomized 88 patients toshort-term (�24 hrs), medium-term (24hrs to 7 days), and prolonged (�7 days)continuous sedation with propofol (n �46) or midazolam (n � 42). In the short-term sedation subgroups, time to extuba-tion and time elapsed until normalizationof the alertness level were significantlyshorter in patients treated with propofol(p � .05). In the medium-term sedationsubgroups, the average sedation time wassimilar in both groups. Recovery timeuntil extubation and time elapsed untilreaching normal alertness levels weresignificantly shorter in patients infusedwith propofol (p � .05). In the long-termsedation subgroups, the mean sedationtime was similar in both subgroups, butrecovery time until extubation and timeelapsed to reach normal levels of alert-ness were significantly shorter in patientsin the propofol group. Although the costof propofol was higher than that of mida-

zolam in all three treatment groups, thelonger ICU stay required with midazolamresulted in postsedation care costs higherthan the costs for the propofol group.These findings indicate that propofol is asedative agent with equivalent safety yethigher clinical effectiveness and bettercost-effectiveness ratio than midazolamin the continuous sedation of critically illpatients (61).

Although these studies indicate thatthe costs of sedation with propofol arelower than those with midazolam in theICU, more studies of this type are neededto further assess the true cost of theseagents.

A more consistent recovery rate wasseen with propofol than with midazolam.For infusions of �4 days, propofol recov-ery time was often related to the durationof sedation, whereas midazolam recoverytime was not. After discontinuation of thedrug, most of the patients receivingpropofol recovered in 1 hr or less,whereas most of the patients receivingmidazolam took from several hours to 10days for similar recovery after deep seda-tion. Propofol has not been comparedwith lorazepam in a clinical trial. Such astudy, however, may be difficult to imple-ment inasmuch as propofol is used forshort-term sedation whereas lorazepamis used in the long term. Side effectsassociated with propofol sedation includehypotension, which is more commonwith rapid dose escalation or iv bolusdoses, bradycardia, and hypertriglyceride-mia, which appears to occur with higherinfusion rates (35).

A new formulation, 2% propofol,which is twice as concentrated as theavailable 1% formulation, is currentlyundergoing end-stage FDA review for useas a sedative in the ICU. The rationale forthis new formulation is that, by doublingthe concentration, the fat load will bereduced by half while maintaining thesame sedative efficacy, thus lessening thelikelihood of increased serum levels oftriglycerides. Ewart et al. (63) conducteda feasibility study comparing 2% propofolwith the 1% formulation in 40 patients(20 in each treatment group) undergoingmechanical ventilation in an ICU aftercoronary artery bypass surgery. No signif-icant differences in the amount of propo-fol used, the rate of infusion, and thenumbers of changes in infusion rate, re-covery time, and time to extubation wasfound between the two formulations.However, mean heart rates of patients


receiving 2% propofol were significantlyhigher throughout the study.

A study conducted by McLeod et al.(64) was designed to determine serumconcentrations of lipids during infusionof 2% propofol for 50 hrs in 30 ventilatedsurgical, trauma, and medical patients inan ICU. The triglyceride concentrationdid not significantly increase over a 50-hrperiod, and both mean cholesterol andhigh-density lipoprotein levels were low.There was a direct correlation betweentriglyceride and C-reactive protein con-centration, and an inverse correlation be-tween cholesterol and C-reactive protein,which suggests that lipid changes in crit-ically ill patients may be in part related tothe acute-phase response. The investiga-tors suggest that, to avoid fat overload incritically ill patients, administration ofadditional lipids be adjusted to accountfor the lipid content of propofol. Somestudies show higher propofol require-ments in first few days of sedation ther-apy with use of 2% propofol. The reasonfor this remains to be determined.

Within 1 yr of the introduction ofpropofol in the United States in 1989,reports appeared of clusters of infectionsin surgical patients who had receivedpropofol (65). This resulted in the inclu-sion of an additive to help retard growthof microorganisms. The additive, ethyl-enediaminetetraacetic acid (EDTA), at aconcentration of 0.005%, has no effect onthe physical or chemical stability of theemulsion components. In the 4 yrs sincethe introduction of this modified propofolpreparation, clinical experience in morethan 30 million patients in the UnitedStates has demonstrated a reduction inthe incidence of fevers and infectionsfrom approximately 20 per year to essen-tially zero (66). EDTA is a chelator ofvarious ions, including calcium. In a pro-spective, randomized, multicenter trial,122 surgical ICU patients requiring ven-tilation were treated with either the orig-inal formulation of propofol or the mod-ified formulation containing EDTA. TheEDTA-containing formulation had no ef-fect on calcium or magnesium homeosta-sis, renal function, or sedation efficacycompared with the original formulation.Of interest was the finding that patientsreceiving the EDTA formulation had asignificantly lower mortality rate at 7 and28 days than those receiving the originalformulation, although this study was notdesigned to evaluate mortality as a pri-mary end point (67). Other potential ef-fects of EDTA also relate to the ability of

this compound to bind cations. TheEDTA-containing formulation of propofolincreases excretion of zinc, which candiminish the inflammatory response tostress by decreasing the release of cyto-kines involved in inflammation, such astumor necrosis factor, and generation offree radicals and other oxidants. How-ever, the implications of these effects re-main to be determined (67, 68).

A generic formulation of propofol hasrecently become available in the UnitedStates. The major differences between theproducts are that the generic formulationcontains a different preservative, sodiummetabisulfite (0.025%), and has a lowerpH (4.5–6.4) to maintain antibacterial ac-tivity of the sulfite than does the EDTAformulation (0.005%), pH (7.0–8.5).

Tests conducted by Redhead et al. (69)compared characteristics of the two for-mulations of propofol. Overall, importantdifferences were found between them,both with respect to physicochemicalcharacteristics and antimicrobial effec-tiveness. In one test, samples of each for-mulation were subjected to excessiveshaking, a well-known test of emulsionstability. After 2 hrs of shaking the ge-neric formulation, the particle-size distri-bution of droplets had changed, and fur-ther changes were observed after anadditional 8 hrs of shaking. In contrast,propofol with EDTA underwent nochanges with 16 hrs of shaking. In a testof stability, the samples were left exposedto air for up to 48 hrs. The genericproduct underwent a pH change of from6.3 to 4.2 and turned to a yellow color,and degradation products were foundby chemical analysis. Propofol withEDTA maintained a constant pH andappearance, and no degradation prod-ucts were detected. Propofol with EDTAslows the growth for at least 24 hrs of awide range of microorganisms, includ-ing those most likely to be found in ahospital. In this study, the effect of thetwo formulations on killing of a widerange of microorganisms was tested.None of the microorganisms grew bymore than 1.0 log unit in 24 hrs in theEDTA emulsion, whereas, in the ge-neric product, one strain each of Esch-erichia coli and Candida albicans grewby more than 1.0 log units in the sametime frame. Despite these differences,the FDA considers the two formulationsto be bioequivalent and interchangeable(i.e., AB rated).

Haloperidol

Haloperidol, a butyrophenone neuro-leptic drug, is the agent of choice fortreatment of delirium in critically ill pa-tients. Clinical effects are observed within30–60 mins after iv administration andlast for as long as 4–8 hrs. The usualstarting dosage is 2–10 mg iv, repeatedevery 2–4 hrs (14). Most patients beingtreated for ICU delirium require muchlarger doses of the drug than noncriti-cally-ill patients (11). Haloperidol doesnot cause major respiratory depression.The drug blocks dopaminergic transmis-sion at postsynaptic receptor sites in thecentral nervous system. Patients treatedwith haloperidol generally seem to bemore calm and are better able to makeappropriate responses (70).

The adverse effects associated withhaloperidol include occasional hypoten-sion resulting from the �-blocking prop-erties of the drug. Although rare with ivadministration, haloperidol may causeextrapyramidal effects such as drowsi-ness, lethargy, a fixed stare, rigidity, andakathisia. These symptoms are usuallymild and reversible with discontinuationof the drug (6, 71). High doses of the drugare associated with QT interval prolonga-tion and development of torsades depointes. The QT interval should be mon-itored closely, and administration of hal-operidol should be discontinued if the QTinterval is prolonged by more than 25%or is �450 msecs (35). Rarely, a patientmay experience neuroleptic malignantsyndrome, a rare complication of halo-peridol therapy with a mortality rate of20% to 30%. Neuroleptic malignant syn-drome develops slowly over 24–72 hrsand can last for up to 10 days after dis-continuation of the drug (72).

Dexmedetomidine

Dexmedetomidine, a selective �-2 ad-renergic receptor agonist, exhibits sym-patholytic, sedative, and analgesic effects,and is eight times more potent for �-2receptor than clonidine. The drug hasbeen approved by the FDA as a short-termsedative (�24 hrs) and analgesic in thecritical care setting, specifically for use inthe early postoperative period (38).

Dexmedetomidine acts at two adren-ergic sites. On the one hand, the drugworks by presynaptic activation of the �-2adrenoceptor, thereby inhibiting the re-lease of norepinephrine and terminatingthe propagation of pain signals. Also, by


postsynaptic activation of these receptorsin the central nervous system, dexme-detomidine inhibits sympathetic activitywith a resultant decrease in blood pres-sure and heart rate. Together, these twoeffects can produce sedation, anxiolysis,sympatholysis, and analgesia (73).

Dexmedetomidine has several advan-tages for use as a sedative in the ICU.Because the drug does not cause respira-tory depression, a patient can be extu-bated without prior discontinuation. Be-cause a dexmedetomidine infusion can becontinued during the postextubation pe-riod, the drug provides flexibility in thetiming of extubation and may be usefulduring the weaning process. Another ad-vantage of the drug is easy arousability oftreated patients—i.e., they can be calmlyand easily awakened (38). The adverseeffects of dexmedetomidine include hypo-tension, hypertension (with the loadingdose), and bradycardia (74).

Two randomized, double-blind, paral-lel, placebo-controlled, multicenter stud-ies evaluated the safety and efficacy ofdexmedetomidine in mechanically venti-lated patients. The starting dose andmaintenance infusion were titrated toachieve mild sedation with arousal to ver-bal commands. In both studies, approxi-mately 60% of patients in the dexmedeto-midine group required no additionalsedation. There were reductions in theneed for supplemental propofol and mi-dazolam of sevenfold and fourfold, re-spectively, compared with placebo recip-ients. In addition, dexmedetomidinereduced the requirement for morphine by50% in both studies (38). Dexmedetomi-dine may lack amnestic properties, how-ever, inasmuch as a small number of pa-tients who received the drug recalledtheir ICU stay and found the experiencevery stressful (73). Because elimination isprimarily hepatic, doses should be de-creased in patients with hepatic dysfunc-tion. Pharmacodynamic responses maybe altered in the presence of both hepaticand renal dysfunction, although no doseadjustment is needed in renal dysfunc-tion (74).

Although promising as a sedativeagent with analgesic-sparing propertiesin the ICU, dexmedetomidine needs to bestudied further with respect to its prop-erties as a sedative and its side-effect pro-file, including studies longer than 24 hrs.For example, the amnestic properties ofthe drug need to be better elucidated.Also, inappropriate use of dexmedetomi-dine might induce or aggravate cardiac

conduction defects or lower cardiac out-put (38). Appropriate patient selection isof the greatest importance, as the hemo-dynamic status of a patient may increasethe likelihood of adverse effects. ICU pa-tients who have hypovolemia, bradycar-dia, or low cardiac output should not betreated with dexmedetomidine (37). Al-though dexmedetomidine may be initi-ated with a loading infusion over 10–20mins, therapy in some patients may beginwith a maintenance infusion that is thentitrated to the desired effect. Dexmedeto-midine is a promising agent with multi-ple actions that reduces analgesic andother sedative requirements and pro-duces a cooperatively sedated patient.Proper patient selection may reduce theincidence of adverse drug events.

MANAGING SEDATIVE AGENTSIN COMMON ICU SETTINGS

The choice of a sedative for intubation,maintenance of ventilation, and extuba-tion profoundly influences outcome, bothin terms of the patient and the economicimpact. Not all patients are candidates fora single sedative agent, and clinicians arefaced with numerous choices when decid-ing which sedative is appropriate for anindividual patient. Older agents, such asthe benzodiazepines, are often the seda-tive of choice; but in recent years, manynew sedatives have become available andthey need to be thoroughly understood inclinical settings.

Sedation in CardiacPostsurgical Patients

The introduction of economic con-straints has encouraged the minimiza-tion of postoperative intensive care. Thisminimization has stimulated interest inearly extubation or “fast track” anesthesiaafter cardiac surgery. Because many crit-ical care nursing standards now require a1:1 nurse-patient ratio for newly venti-lated postoperative patients, early extuba-tion may reduce nursing requirements orallow the patient to be transported to lessintensive care areas (75).

The choice of a sedative in this patientpopulation is a major determinant of out-come. The appropriate agent would haverapid onset of action, speed and ease ofdose titration, rapid recovery from seda-tion with fast weaning and short time toextubation, hemodynamic stability dur-ing maintenance of sedation, and controlof stress responses while maintaining ad-

renocortical responsivity to adrenocorti-cotropic hormone stimulation. Two earlystudies by Grounds et al. (50) andMcMurray et al. (54) compared sedationwith propofol and midazolam in 160 pa-tients who had undergone coronary ar-tery bypass surgery (CABS). The resultsof the trial indicated that propofol per-mitted a significantly faster time to extu-bation than the other sedatives studied(Figs. 1 and 2).

In another study by Roekaerts et al.(76), continuous infusions of midazolamand propofol were compared after coro-nary artery surgery in 30 patients whounderwent deep sedation for a mean of 9to 10 hrs. There was no difference in thequality of sedation between the two treat-ment groups, but patients treated withpropofol had a faster recovery from deepsedation and faster weaning from theventilator. Ostermann et al. (30) recentlypublished a systematic review of random-ized trials comparing sedatives in the ICUsetting. Of eight trials that examined therelative effectiveness of propofol and mi-dazolam for time to extubation in post-cardiac surgery patients, five found thatthis time was shorter for propofol thanfor midazolam.

Two large, prospective, randomizedstudies compared the efficacy and safetyof early and conventional extubation.Cheng et al. (77) conducted a prospective,randomized, controlled clinical trial,evaluating morbidity outcomes and safetyof a modified anesthetic technique to pro-vide shorter sedation and earlier extuba-tion times (1–6 hrs) than those of a con-ventional anesthetic protocol used forprolonged sedation and extubation(12–22 hrs) in 120 patients after CABS.This trial demonstrated that early tra-cheal extubation is safe in this patientpopulation and does not increase periop-erative cardiac, respiratory, hemody-namic, or sympathoadrenal morbidity.The postextubation intrapulmonaryshunt fraction was improved, and boththe ICU and hospital LOS were reduced(Fig. 3).

In the early extubation group, anes-thesia induction consisted of 15 �g/kgfentanyl � 50 mg thiopental. Anesthesiawas maintained with isoflurane beforesurgery. A propofol infusion at 2–6 mg/kg/hr was commenced at the start of sur-gery and maintained until 1–4 hrs in theICU. In the conventional extubationgroup, anesthesia induction consisted of50 �g/kg fentanyl. A 0.1-mg/kg injectionof midazolam was administered in the


prebypass period. Isoflurane was used asrequired during the perisurgical period.In the ICU, routine infusions of morphine(2–10 mg/hr) and midazolam (1–3 mg/hr) were adjusted to achieve the samedegree of sedation as in the early extuba-tion group. Fifty-one of the 60 patients ineach group (85%) were extubated withinthe defined time period. Postoperative ex-tubation time and ICU and hospitallengths of stay were significantly shorterin the early extubation group. At 48 hrsafter operation, no significant differencewas found between the two groups inpostoperative myocardial ischemia inci-dence and ischemia burden, creatine ki-nase-MB levels, plasma catecholamine(all within the normal clinical range),and ventilatory morbidity. Postextuba-tion apnea characteristics and incidencesand degree of atelectasis were similar be-tween the groups. Intrapulmonary shuntfraction improved significantly in theearly group at 4 hrs after extubation.There was a similar incidence of treatedpostoperative complications in the twogroups, but three patients in the conven-tional extubation group died of stroke orpostoperative myocardial infarction (77).

In another randomized controlledtrial conducted by the same investigators,the costs of therapy for early and lateextubation, and the time parameters forICU and hospital stay, were compared inpatients after CABS. Early extubation sig-nificantly reduced the cost of coronaryICU stay by 53% (p � .026) and the totalcost of CABS by 25% (p � .019) whencompared with late extubation. In eachgroup, 41 of 50 patients (82%) were ex-tubated within the defined period, andboth the ICU LOS and the overall hospitalLOS were significantly lower for the earlyextubation group (p � .046 and p � .015,respectively) (78) (Fig. 4).

Because late extubation and conven-tional anesthesia for CABS are well-

established practices, the major modifica-tions required for early extubation shouldbe thoroughly evaluated and should in-clude postoperative intensive care man-agement. Through the use of an appro-priate anesthetic technique andpostoperative management, Silbert et al.(75) demonstrated that early extubationcan be achieved after CABS without ma-jor complications. In a prospective, ran-domized, controlled trial, 100 patientsundergoing elective CABS were random-ized to early extubation or conventionalextubation. Those in the early extubationgroup received a reduced dose of fentanyl(15 mg/kg) and an anesthetic compatiblewith early extubation, whereas those ran-domized to conventional extubation re-ceived fentanyl at a dose of 50 mg/kg. Inthe early extubation group, anesthesiawas augmented by administration ofpropofol for induction and maintenance.The median time to extubation in theearly extubation group (240 mins) wassignificantly less than that in the conven-tional extubation group (420 mins) (p �.01). Importantly, early extubation didnot result in an increased rate of reintu-bation, postoperative myocardial infarc-tion, or other complications. The authorsnoted that, besides demonstrating thatearly extubation is as safe as conventionalextubation, there are several theoreticaladvantages of the technique, includingearlier mobilization of the patient, de-creased risk of nosocomial infection, bet-ter pulmonary function and improved he-modynamics. They pointed out, however,that it is not known if these “benefits”will prove significant in practice.

Delirium in the ICU

Delirium, a common disorder in ICUpatients, has often been referred to as

“ICU psychosis.” This term for uncon-trolled agitation is, however, inappropri-ate and nonspecific (35). It was intro-duced to underline the etiologicalsignificance of psychosocial and psycho-logical factors in understanding the syn-drome (5). A more specific definition ofdelirium is “an acute, reversible organicmental syndrome with disorder of atten-tion and cognitive function, increased ordecreased psychomotor activity, and adisordered sleep-wake cycle.” The esti-mated prevalence of delirium in the ICUis 15% to 40% and is escalating as aresult of increases in the number of el-derly and more severely ill patients ad-mitted to the ICU. In this setting, delir-ium contributes to increased morbidityand is associated with a poorer prognosisand a mortality rate of 10% to 33% (79).

Delirium is a consequence of a non-specific central nervous system reactionto disruption of the internal environmentthat is necessary for normal function.Predisposing factors for delirium includeadvanced age, underlying primary cere-bral illnesses such as dementia and Alz-heimer’s disease, and a history of alcoholor substance abuse. Underlying chronicsystemic illness accentuated by metabolicand hemodynamic instability, hypoxemia,acidosis and electrolyte imbalances, se-vere infections, and intracerebral abnor-malities, such as brain tumors, can alsoprecipitate delirium. ICU-related factorscontributing to the development of delir-ium include sleep deprivation, sensoryoverload, lack of meaningful verbal orcognitive stimulation, and immobiliza-tion. Withdrawal of drugs such as opioids,sedatives, and several other pharmaco-logic agents can also contribute to thedevelopment of delirium (79). Among themore common causes of altered mentalstatus in critically ill patients are adversedrug reactions and drug–drug interac-tions. Numerous drugs, including thosewith anticholinergic properties, cardio-vascular drugs, H2-receptor antagonists,and antimicrobials all can be responsiblefor mental disturbances and delirium(80).

The differential diagnosis of deliriumincludes dementia, depression, andschizophrenia. Dementia develops slowlyand is long lasting, whereas delirium hasan acute onset and recovery is almostalways complete. The hypoactive form ofdelirium may be mistaken for depression,but disorientation, which is common indelirium, is not a feature of depression.Acutely schizophrenic patients may seem

Figure 3. Time to tracheal extubation in 30 pa-tients sedated with propofol and 30 patientstreated with midazolam who received mechanicalventilation in an intensive care unit after cardiacsurgery. ‡Range 5–7; §range 80–610; p � .001.

Figure 4. Time to tracheal extubation in 50 pa-tients sedated with propofol and 50 patientstreated with midazolam who received mechanicalventilation in an intensive care unit after coro-nary revascularization. ‡Range 2.9–19.1; §range73.6–208.5; p � .001.


confused, but examination reveals thatthey do not have cognitive deficits. Inaddition, schizophrenia is associated withauditory, rather than visual, hallucina-tions (79).

Two distinct clinical presentations ofdelirium have been observed. In hyperac-tive delirium, patients are restless andagitated; conversely, those with the hypo-active variant exhibit decreased con-sciousness and psychomotor activity. Amixture of hyperactive and hypoactive de-lirium is also seen in some patients (79).An interesting characteristic of deliriumis that the behavior of the patient canchange dramatically within hours or evenminutes. Drowsiness and lethargy canchange to alertness and lucidity for atime, and then can quickly change toagitation and aggression (81).

The delirious patient sometimes in-correctly perceives the environment ashostile or threatening. The patient mayattempt to escape, necessitating the useof physical or chemical restraints, or maytry to assault staff and visitors. There isalso an increased risk of self-harm result-ing from unintentional dislodgment ofcritical life-support and monitoringequipment. Such a situation often pro-longs the length of an ICU stay, necessi-tates further invasive treatment, andincreases the risk of additional complica-tions. Also important to note is that, be-cause of the impairment of short-termmemory associated with delirium, a pa-tient may not even remember an episodeof delirium once it has subsided (79).

If a patient exhibits unsafe behavior,insomnia, hallucinations, delusions, agi-tation, or psychomotor hyperactivity,pharmacologic therapy should be consid-ered (14). In most ICUs, a neurolepticagent is the recommended medication fortreatment of delirium resulting fromcauses other than withdrawal. Haloperi-dol is generally the neuroleptic agent ofchoice because, in addition to its efficacy,this drug has few anticholinergic and hy-potensive effects. Other agents that areoften used to sedate patients and enhancesleep—including benzodiazepines, anti-histamines, and hypnotics— usuallyworsen delirium (81). Nevertheless, ben-zodiazepines are occasionally given incombination with haloperidol, which al-lows for the use of smaller and safer dos-ages of either agent alone (79). Doses ofmedications used to treat a patient’s pri-mary condition should be reduced or dis-continued if they contribute to delirium.If the drug cannot be discontinued, a

change to a similar drug with less risk ofdelirium is advisable. Analgesics shouldalso be given if the patient is experiencingpain. Patients with delirium related toalcohol or drug withdrawal may continueto be delirious even when their with-drawal symptoms are being adequatelytreated. In these situations, neurolepticagents should be added to the medica-tions specified in a withdrawal protocol(81).

Respiratory Failure and Patient-Ventilator Asynchrony

Patients undergoing mechanical ven-tilation are likely to breathe out of syn-chronization with the ventilator when ag-itation resulting from fear and anxietycauses tachypnea (26). Changes in thepatient’s respiratory status and the devel-opment of asynchrony between the pa-tient and the ventilator may also repre-sent a possible emergency situation.“Fighting” or “bucking” the ventilator de-scribes the presence of agitation and re-spiratory distress in the ventilated pa-tient. Because agitation leads to anincrease in CO2 and lactic acid produc-tion, life-threatening respiratory andmetabolic acidosis may occur. This de-synchronization between efforts of inspi-ration and their rhythm with the ventila-tor can result in ineffective oxygendelivery and CO2 elimination. Some ofthe signs of respiratory distress are tachy-pnea, diaphoresis, and cardiovascular ab-normalities (82).

There are numerous possible causes ofsudden respiratory distress. Ventilator-related causes include improper settingof the ventilator and malfunctions of theequipment. Causes related to the airwayinclude malposition of the endotrachealtube, cuff problems, endotracheal ob-struction, and airway trauma from tra-cheostomy tubes. Patient-ventilator asyn-chrony may be caused by inappropriateventilator selection or settings, inade-quate FIO2 or positive end-expiratorypressure level, and ventilatory rate. Fi-nally, causes related to the patient in-clude abnormalities in the airway, lungparenchyma, and pleural space, as well ascardiovascular dysfunction and alteredventilatory drive (83).

Appropriate sedation is especially im-portant in patients with respiratory fail-ure. When sufficient doses are adminis-tered, sedatives can diminish patientstruggle against mechanically supportedbreaths, improve chest wall compliance,

and allow manipulation of inspiratory toexpiratory ratio and other ventilator vari-ables to maximize oxygenation (84). Ifhypoxia caused by circulatory failure isthe indication for mechanical ventilation,propofol or midazolam, which can affectsystemic vascular resistance, should becarefully and slowly titrated. Respiratoryfailure developing from other causes ofhypoxia is not exacerbated in the sedatedpatient if care is taken to insure properventilation with appropriate delivery ofoxygen concentration. In addition to itssedative properties, propofol, in contrastto other agents used for sedation, may bebeneficial to patients with severe air flowobstruction; studies have demonstratedthat this agent reduces pulmonary resis-tance in ventilated chronic obstructivelung disease patients (26).

Sedation During Weaning fromMechanical Ventilation

Managing agitation and pain in me-chanically ventilated patients who areready for weaning requires a thoroughunderstanding of the available pharmaco-logic agents, because their manifesta-tions can profoundly influence the out-come of weaning. It is now well knownthat patients being weaned from mechan-ical ventilation require appropriate seda-tion for a successful outcome with re-spect to extubation and release from theICU. The stresses of the ICU environ-ment, including bright or flashing lights,alarms, hectic pace, and exposure to un-familiar personnel, often lead to anxietyand agitation. In addition, sleep disrup-tion, undergoing numerous tests andprocedures, immobility for extended timeperiods, and physical restraints all fur-ther necessitate the need for sedation(71).

Nonpharmacologic intervention at thetime of weaning may relieve mild anxiety.Such interventions include changing theenvironment, using relaxation tech-niques, reassuring the patient, and pro-viding adequate rest and psychologicalsupport. However, for patients who donot respond to these interventions, phar-macologic therapy should be instituted,and sedatives should be given on a regu-larly scheduled basis to promote stableblood levels (71).

Agents that can contribute to signifi-cant respiratory depression should beavoided when a patient is being weanedfrom mechanical ventilation. Opioids andbenzodiazepines should not be used or, if


the patient is already being treated withthese agents, they should be discontinuedor reduced during the weaning process.However, because patients often becomemore highly anxious during weaning,there is a real need for sedation. Haloper-idol, a neuroleptic, is often employed dur-ing the weaning process because it doesnot produce respiratory depression.Propofol is another useful drug for seda-tion during the weaning process because,compared with benzodiazepines, it has aquick onset and short duration of action,thereby reducing the time needed for re-covery of spontaneous respiration (85).

Several studies have attempted cost-benefit analyses by comparing propofoland midazolam for sedation in patientsreceiving mechanical ventilation in theICU. In a study by Carrasco et al. (61),critically ill patients were allocated to re-ceive short-term (7 days) continuous se-dation with either midazolam or propo-fol. Propofol was more expensive thanmidazolam, but there was a cost savingsof approximately $18 per patient in thepropofol group that was attributable to ashorter ICU stay. Barrientos-Vega et al.(59), in an open-label, randomized, pro-spective trial, compared the effectivenessof sedation, time required for weaning,and costs of prolonged sedation of criti-cally ill patients undergoing mechanicalventilation for more than 24 hrs withmidazolam or propofol. Midazolam andpropofol were equally effective as sedativeagents. Despite large differences in thecost of the two agents for sedation, theeconomic profile was more favorable forpropofol than for midazolam because ofthe shorter weaning time for patients re-ceiving propofol. On average, the mida-zolam group required �4 days to awakenand wean from mechanical ventilatorysupport once the infusion was termi-nated, whereas the propofol group aver-aged 35 hrs (p � .0001).

In a multicenter, randomized, open-label trial, Hall et al. (51) comparedpropofol and midazolam, given for differ-ent durations of time, on extubation timeand LOS in the ICU in 99 evaluable crit-ically ill patients (53 in the midazolamgroup and 46 in the propofol group) infour different types of ICUs. After admis-sion to the ICU, physicians assessedwhether patients would require sedationfor short-term (�24 hrs), medium-term(�24 hrs and �72 hrs), or long-term(�72 hrs) mechanical ventilation. Thedose of each drug was adjusted to achievea daily-targeted Ramsay Sedation Scale

score. Sedation with propofol was associ-ated with a shorter time to tracheal ex-tubation than sedation with midazolam,but there was either no difference in thetime to ICU discharge or a prolongedtime for the propofol group. The authorsspeculated that this difference with re-spect to time to discharge from the ICUmight be accounted for by a delay inpatient transfer secondary to systematichandling of the patients or, alternatively,patients in the propofol group may haverequired more ICU care for other criticalillnesses.

Walder et al. (86) conducted a system-atic review of 27 randomized trials toestablish the efficacy and harm of propo-fol vs. midazolam in mechanically venti-lated patients. In 13 trials, mostly post-operative, sedation lasted from 4 to 35hrs. In nine of these trials, the averageweaning time from ventilation was 0.8–4.3 hrs with propofol and 1.5–7.2 hrswith midazolam. There was a relativelyshorter weaning time with propofol in sixtrials and with midazolam in one trial; inone trial, the time was equivalent for thetwo drugs. Across all trials, the adequacyof sedation with propofol was longer thanwith midazolam. The authors suggestedthat propofol, because of its rapid redis-tribution compared with midazolam, maybe advantageous when frequent dose ad-justments are required, such as in agi-tated patients. There was also strong ev-idence that weaning times were shorterafter sedation for �36 hrs with propofol.

In an interesting approach to improv-ing on the sedative effects of both mida-zolam and propofol, and to take advan-tage of the best features of each drug,studies have been conducted to evaluatetheir combined use. The interaction be-tween propofol and midazolam is syner-gistic rather than simply additive, asdemonstrated in a prospective, con-trolled, randomized, double-blind trialconducted by Carrasco et al (61). Thecombination of the two agents was com-pared with each agent alone in post-CABSpatients. Combined therapy was equallyas effective as either agent alone and wasassociated with rapid awakening and ex-tubation, reduction in overall sedativedosage, and resultant lower pharmaceu-tical acquisition cost. This study high-lights an interesting and potentially use-ful drug interaction between midazolamand propofol and offers a promising areaof investigation for future studies of ICUsedation (27).

USING SEDATIVE AGENTS INSPECIAL ICU CIRCUMSTANCES

The care of patients in the ICU ishighly challenging, not least because ofdifferences between patients that can sig-nificantly affect the outcome of manage-ment. Age, personal characteristics, un-derlying disease, and the nature of theinsult leading to admission to the ICU allprofoundly affect the decision-makingprocess for patient management. Seda-tion is a key part of treatment in the ICU;patients adapt more easily to intubationand mechanical ventilation when they re-ceive the appropriate sedative and painmedication. Sedation must be individual-ized to the patient. Benzodiazepines maybe appropriate for one patient, whereaspropofol may be preferred for another.Following is a discussion of some of thespecial groups of patients who require aspecific approach to sedation—alcoholand drug abusers experiencing with-drawal symptoms when confined to theICU, patients with status asthmaticus, pa-tients undergoing end-of-life terminalweaning, pregnant women, patients un-dergoing endotracheal intubation, andtraumatic head injury patients.

Many critically ill patients fall intospecial clinical situations that must betaken into consideration when institutingsedation. The critical care practitioner isfrequently challenged in the ICU by theinability to quickly and easily diagnosesituations, such as drug or alcohol with-drawal, that may interfere with the in-duction of sedation. Withdrawal syn-dromes in individuals with a history ofheavy alcohol and benzodiazepine usehave historically been associated withhigh rates of morbidity and mortality.

Alcohol Withdrawal

Delirium tremens is the most seriousmanifestation of the alcohol withdrawalspectrum. It is seen in approximately 5%of hospitalized patients with a history ofalcohol abuse, and has a mortality rateranging from 1% to 15% (64). With-drawal symptoms can progress over a pe-riod of 24–72 hrs to delirium tremens, acondition marked by agitation, tremor,and an acute state of confusion associatedwith disorientation, hallucinations, andautonomic hyperactivity. Whenever pos-sible, treatment of alcohol withdrawalshould be initiated before the onset ofagitated delirium. Patients with worsen-ing conditions and those with concomi-


tant medical problems require admissionto the ICU. Here, control of seizures,maintenance of hemodynamic stability,arrhythmia management, airway protec-tion, and correction of nutritional andmetabolic deficiencies are facilitated withinitiation of pharmacologic therapy forwithdrawal (37).

Once heavy alcohol use has been iden-tified, proper prophylaxis should be insti-tuted, both by maintaining optimal elec-trolyte levels through potassium,magnesium and phosphorous replace-ment, and by administration of thiamine,vitamin B12, and folate together with anappropriate sedative. The most importantpharmacologic treatment is use of agentsthat are cross-tolerant with alcohol,thereby providing prophylaxis against sei-zures and relieving the frequently intenseagitation, hallucinosis, and tremulous-ness. Although treatment with alcohol iseffective and can be intravenously ti-trated, such treatment is not addressed inmost reviews and has not been well stud-ied in clinical trials. The short duration ofaction of ethanol requires prolonged ad-ministration and does not always elimi-nate the need for additional therapy. Themost widely administered pharmacologicagents for the treatment of alcohol with-drawal are benzodiazepines (37).

Alcohol ingestion affects many regula-tory systems; among the consequencesare an increase in the release of endoge-nous opiates, activation of the GABA-Areceptor, inhibition of the N-methyl-D-aspartate (NMDA) receptor, and interac-tions with serotonin and dopamine recep-tors. Chronic exposure to the inhibitoryGABA-A and excitatory NMDA receptorsis believed to play a role in the pathogen-esis of alcohol withdrawal. The long-termeffects of alcohol on the number andfunction of central nervous system recep-tors cause excessive central nervous sys-tem excitability during periods of absti-nence, resulting in the signs andsymptoms of delirium tremens (87).

Treating alcohol withdrawal usuallyincludes the substitution of an agent witheffects on the GABA-A receptor. Becausebenzodiazepines potentiate this neuro-transmitter, they have been successfullyused to reduce the signs and symptoms ofwithdrawal. Barbiturates are not recom-mended because they have a narrow ther-apeutic index, and haloperidol is less ef-fective in preventing delirium andseizures. Propofol may be an alternativeto benzodiazepines for controlling alco-hol withdrawal symptoms, but there is

limited data available supporting its usein nonintubated patients (35).

With the exception of a few case re-ports, there have been limited studies ofpropofol for treatment of patients withdelirium tremens in the ICU setting (87,88). There are, however, several proper-ties of propofol, including less cross-tolerance than traditional benzodiaz-epines, ease of titratability, and a rapidmetabolic clearance, that make it a prom-ising drug for sedation in patients withsevere alcohol withdrawal and deliriumtremens. Like alcohol, propofol affectsboth the GABA-A and glutamate recep-tors (87).

Other Withdrawal andIntoxication Syndromes

Another complicating factor in themanagement of ICU patients is the pres-ence of symptoms related to either thewithdrawal of drugs or drug intoxicationresulting from adverse effects, or drug–drug interactions. These symptoms mayarise with drugs used therapeutically inthe ICU, or with licit or illicit drugs thatthe patient used before hospital admis-sion.

Withdrawal syndromes are a frequentoccurrence in the ICU, especially in ur-ban locations, because 36% of intentionalinjury victims are drug-dependent. With-drawal syndromes confuse the clinicalmanagement of such patients and may beextremely difficult to diagnose. Thesesyndromes are often lethal, and prophy-lactic measures should be taken to pre-vent their emergence in all patients iden-tified at risk. Therefore, it is safe toconsider all ICU patients to be at highrisk for drug or alcohol dependence un-less proven otherwise. Where there isdoubt, patients should be tested for evi-dence of drugs and interviewed togetherwith family members for the presence ofdrug- dependence traits. Appropriate pa-tients should be referred for formal eval-uation and treatment once they havebeen stabilized. Withdrawal syndromesmust be promptly recognized, differenti-ated from traumatic or metabolic deteri-oration, and treated. The mainstay ofmost withdrawal therapy is supportivecare and treatment with the appropriatesedative (i.e., benzodiazepines or propo-fol). In consideration of the high rate ofmultiple intoxicants present in traumapatients, withdrawal can occur from mul-tiple agents in a single patient, further

compounding the difficulties inherent inmanaging this patient population (81).

Withdrawal from benzodiazepines inthe ICU includes an abstinence syn-drome, which is marked by anxiety, fear,confusion, and agitation. In addition, thepossibility of tachycardia and panic at-tacks may occur as the patient emergesfrom sedation. Severe withdrawal symp-toms, including refractory seizures, maybe seen when benzodiazepines are dis-continued in critically ill patients whohad been receiving treatment with theseagents before their hospital admission.Treatment with a benzodiazepine such asoral lorazepam is appropriate for absti-nence or withdrawal symptoms, withslow tapering of the dose. Intravenousagents such as lorazepam or midazolamcan be used in intubated patients.Clonidine and beta-adrenergic blockerscan be administered to modify symptomsand improve tolerance to benzodiazepinewithdrawal (89).

Narcotic withdrawal is common in pa-tients receiving long-term therapy withopioids for palliative care of cancer orchronic pain syndromes, as well as inpatients with a history of narcotic abuse.Replacement of the narcotic with contin-uous infusions of fentanyl or morphinesulfate, or administration of methadone,is commonly used in the ICU.

Cocaine, a sympathetic-stimulatingdrug, increases the release of presynapticnorepinephrine and blocks its reuptake.This action causes various cardiopul-monary and neuropsychiatric effects in-cluding tachycardia, hypertension, respi-ratory depression, anxiety, tremor,seizures, and hyperthermia (90). Wheninitiating sedation in patients who are ina hypercatecholaminic state, it is impor-tant to determine both the patient’s his-tory of cocaine use and evidence of with-drawal symptoms. Benzodiazepines arecommonly used for sedation of patientswith suspected or known cocaine abuseand, although propofol has occasionallybeen administered, its use for sedationof cocaine abusers is not strongly sup-ported in the literature. The hypercat-echolamine state must also be treated inthese patients; both �- and �-blockershave been used successfully. In addition,all patients suspected of a withdrawalsyndrome should be rapidly evaluated forother physiologic causes such as hypox-emia, hypercarbia, or electrolyte abnor-malities.

Padula and Willey (91) examined thehypothesis that smokers undergoing


forced abstinence from tobacco in a car-diac ICU would be more anxious thannonsmoking patients and exhibit morewithdrawal symptoms. There were 16smokers and 17 nonsmokers enrolled inthe study. The investigators evaluatedtwo types of anxiety, state anxiety (a mea-sure of situational anxiety) and trait anx-iety (a measure of general anxiety). Thepresence of withdrawal symptoms wasbased on patient perception of increasedheart rate, degree of calmness, and de-gree of restlessness. The results of thestudy indicated that smokers exhibitedsignificantly higher trait anxiety com-pared with nonsmokers, but there was nodifference in state anxiety between thegroups. Neither group reported physicalwithdrawal symptoms, but smokers expe-rienced more psychological withdrawalsymptoms than nonsmokers on the firstday after admission. Reversal of symp-toms can be achieved by using nicotinepatches, which are commonly used inpatients with multiple traumas.

Status Asthmaticus

Patients presenting with status asth-maticus, or severe asthma that is unre-sponsive to standard therapy, usually re-quire mechanical ventilation andsedation until respiratory function im-proves. Benzodiazepines are the mostcommonly used sedative in these pa-tients. Propofol may be most appropriatefor asthmatic patients inasmuch as it hasrecently been shown to have substantialbronchodilatory properties at high dosesnot demonstrated with other sedatives oranalgesics (26). These bronchodilatoryproperties were demonstrated in a studyshowing that propofol reduces pulmo-nary resistance (decreases in airway resis-tance and intrinsic positive end-expira-tory pressure) in patients with chronicobstructive pulmonary disease who wereundergoing mechanical ventilation (92).Other clinical studies have shown a sim-ilar effect of propofol in patients withchronic obstructive pulmonary disease(93). Propofol containing EDTA is com-monly used in patients at risk of statusasthmaticus for its bronchodilatory prop-erties and its lack of a trigger in extrinsicasthma or in patients with sulfite intol-erance (26).

Benzodiazepines have no intrinsicbronchodilating properties, and pro-longed effects from continuous infusionof these agents have been associated withan increase in complications as well as

prolonged ventilator use, ICU and hospi-tal stays. Other agents with bronchodila-tor properties such as ketamine and halo-thane have undesirable side effects (93). Apotential problem with the use of propo-fol is that high doses are required to elicita smooth muscle relaxant effect, raisingconcern about hypotension as an adverseeffect. Little is known about whetherbronchodilation occurs at standard dosesof the drug. However, in patients under-going mechanical ventilation who areshowing high peak pressures and severebronchospasm, the use of propofol in ad-dition to standard therapy for bronchos-pasm may have additional benefits.

Terminal Weaning. In recent years,there has been a greater awareness of theimportance of providing maximum com-fort to terminally ill patients who arebeing weaned from mechanical ventila-tion. In this context, patient comfort isdirectly related to the choice of sedative.

The management of patients undergo-ing end-of-life care in the ICU includes,in many cases, terminal weaning frommechanical ventilation. Of prime impor-tance is that critical care practitionersprovide quality end-of-life care. Once thepatient, the family, and the primary-carephysician have made the decision, theattending physicians in the ICU are re-sponsible for providing the patient a com-fortable, anxiety-free withdrawal frommechanical ventilation.

When a decision to forgo treatment ismade, the focus should be on specifyingthe goals of patient care and assessingtreatments in light of these goals. Theuse of appropriate palliative measures cannearly always control symptoms accom-panying withdrawal of life support. AfterICU interventions are discontinued, pa-tient comfort becomes the most impor-tant objective. This must be assessed fre-quently, and signs of discomfort shouldbe treated with adequate doses of seda-tives and opioids. If terminal weaning ischosen, a limited time course should beagreed on to prevent prolongation of thedying process (94).

Dyspnea and anxiety should be antic-ipated when ventilator support is with-drawn. Opioids and benzodiazepines orpropofol have become the drugs of choiceto treat dyspnea and anxiety or agitation,respectively. These agents should be im-mediately available and titrated to effect,but may also be given before ventilatorwithdrawal to prevent anticipated symp-toms and signs of distress from occur-ring. There is wide variation in the doses

required to relieve symptoms because ofprevious drug exposure, level of toler-ance, drug metabolism, and degree ofawareness. Occasionally, opioid-tolerantpatients require higher doses of mor-phine (94). When choosing a sedative, itis necessary to balance the beneficial ef-fects in terms of patient comfort withpossible toxic effects that may adverselyaffect the respiratory or cardiovascularstate of the patient, thereby increasingthe discomfort level and possibly causingpremature death.

Wilson et al. (95) conducted a study todetermine why and how sedatives andanalgesics are ordered and administeredduring the withholding and withdrawal oflife support. In a total of 22 critically illpatients from each of two ICUs, theyfound that large doses of sedatives andanalgesics were ordered primarily for re-lief of pain and suffering during the with-holding and withdrawal of life support,and that the time to death was not de-creased by drug administration. Thestudy found that, after the initiation ofthe withholding or withdrawal of life sup-port, the median time until death was 3.5hrs in patients receiving drugs and 1.3hrs in those not receiving drugs. Thereasons for drug administration were todecrease pain (88% of patients), anxiety(85% of patients), and air hunger (76% ofpatients); to comfort families (82% of pa-tients); and to hasten death (39%), al-though hastening death was never theonly reason cited. Not surprisingly, sig-nificantly lower amounts of benzodiaz-epines and opiates were given in the 24hrs before withholding and withdrawal oflife support than were given after with-drawal was initiated.

In a Canadian retrospective cohortstudy, Hall et al. (96) compared the use ofsedation and pain relief to prevent andtreat discomfort during the dying processin the end-of-life care of ICU patients whowere or were not withdrawn from lifesupport. In the final 12 hrs of life, therewas a wide variability (greater than ten-fold) among physicians in the two ICUsstudied with respect to prescribed dosesof morphine and sedative agents, whetheror not life support was withheld or with-drawn. Diazepam and midazolam wereused more frequently than lorazepam orpropofol. Doses of morphine and loraz-epam were fivefold higher in patientsfrom whom life support was withdrawn incomparison with patients for whom lifesupport was continued. The amount ofmorphine used in patients withdrawn


from life support increased over the 12-hrperiod and particularly in the final 4 hrsof life. Similar results were noted for theuse of lorazepam, midazolam, and propo-fol.

Pregnancy

Although opioids are known to crossthe placenta and have an effect on neo-natal outcome, there is also evidence thatboth propofol and the benzodiazepinescan crossover to the fetus. Therefore, it isimportant that these drugs be titrated toappropriate levels when used to sedatethe pregnant patient in the ICU. Withrespect to teratogenicity, although it isknown that long-term opioid use affectsintelligence and other neurologic factorsin neonates, there is little evidence forsuch effects with propofol.

It has been reported that several clin-ical trials have evaluated the effect ofpropofol and other sedatives or anesthet-ics on pregnancy outcome in women un-dergoing an assisted reproductive tech-nique. In a multicenter retrospectivepilot trial and survey, Beilin et al. (97)evaluated the effect of sedatives and an-esthetics on pregnancy outcome after ga-mete intrafallopian transfer, a type of as-sisted reproductive technique usuallyperformed laparoscopically under generalanesthesia. Participating in the surveywere seven US fertility clinics represent-ing 455 procedures. The clinical preg-nancy rate (number of pregnancies/number of procedures) was 35% and thedelivery rate (number of women who de-livered at least one live baby/total numberof procedures) was 32%. There was nostatistical difference in either rate be-tween women who received a sedative oranesthetic (propofol, nitrous oxide, mida-zolam, or isoflurane) and those who didnot. Two other clinical studies evaluatedthe effect of propofol on outcome inwomen undergoing assisted reproductivetechnology. In these studies, as in thestudy by Beilen et al., the effect of propo-fol on oocytes was evaluated, and no del-eterious effect was found. In anotherstudy, in which propofol had a negativeeffect on pregnancy rate, the effect of thedrug on embryos, rather than oocytes,was evaluated.

Endotracheal Intubation

In the ICU, a key aspect of endotra-cheal intubation of the patient with fail-ing respiratory status is the ability to

make the patient comfortable rapidlywhile managing the airway. When choos-ing the most appropriate sedative agentfor this purpose, the hemodynamic sta-bility and volume status of the patientmust be considered. Propofol is an appro-priate agent if the volume status of thepatient is acceptable and the patient isnot hypovolemic. In patients who are he-modynamically unstable, however, eto-midate is commonly used since it hasfewer vasodilatory and myocardial de-pressant effects in critically ill patientsthan do other sedative agents. Midazo-lam, thiopental, or methohexital are notused because they can cause hypotension.

Head Trauma

Patients with head injuries present achallenge in the ICU that differs fromtrauma without central neurologic in-volvement. The aim of therapy is to ame-liorate the effects of the initial injurywhile preventing secondary injury suchas edema, infection, and ischemia. Pa-tients with head trauma should always bemonitored by means of neurologic exam-ination. The confusion and agitation re-sulting from brain injury often cause thepatient to struggle and resist nursingcare and mechanical ventilation. Becauseintracranial hypertension is a frequentoccurrence, the effects of sedatives oncerebral metabolism and intracranialelastance must be considered before usein this patient population (37). Elevatedintracranial pressure (ICP) is the mostimportant pathophysiology resultingfrom head injury. An increase in ICP de-creases the cerebral perfusion pressure(CPP), which is the driving force behindcerebral blood flow. Thus, patients withhead injury experience reduced cerebralblood flow (25).

Head injury produces multiple sys-temic effects that must be considered in-patient management. For example, hypo-tension is common after injury to thehypothalamus, brain stem, or spinal cord,so that ablation of the remaining sympa-thetic drive with pharmacologic sedationmay lead to sudden and occasionally se-vere cardiovascular collapse, whichleads to further brain ischemia. Alter-natively, a frequently observed hypera-drenergic state requires that sedationprovide protection from additionalstress, yet not risk the critical carestandard of maintaining organ perfu-sion. Therefore, before treating withsedatives, it is essential to evaluate the

volume status and hemodynamic re-serve of the patient (37).

It is of critical importance that theventilated patient with head trauma beeasily and quickly awakened on a periodicbasis for neurologic evaluation. Afterphysiologic abnormalities have been eval-uated, the degree of discomfort should beaddressed, because pain aggravates thestress response, leading to further in-creases in ICP. Opioids relieve pain andalleviate the hyperadrenergic state byproviding analgesia and sedation, andthese agents normally do not perturb in-tracranial dynamics when ventilation iscontrolled. A disadvantage of the use ofopioids is that these drugs may cloud theneurologic evaluation. Other limiting ef-fects of opioid sedation are the potentialfor gastrointestinal hypomotility and thedelay of the weaning process and success-ful extubation (37).

The aims of treatment for patientswith head injury are reduction and/ormaintenance of ICP within acceptableranges, maintenance of adequate CPP,and minimization of brain activity. Me-chanical ventilation and drug therapy areused to accomplish these aims. Adminis-tration of sedative agents is an integralpart of the management scheme (all sed-atives cause cerebral depression to someextent). The ideal sedative reduces ICPwhile maintaining an adequate CPP. An-other property of an ideal sedative forpatients with head injuries is titratabilitymanagement with diuretics and antihy-pertensive agents that may affect intra-vascular volume. Benzodiazepines havebeen commonly used in this setting,with midazolam being the drug ofchoice because of its short half-life, butthis agent may exhibit a prolonged du-ration of sedation in patients receivingcontinuous infusions for 24 hrs orlonger, with emergence delayed for 1–2days or longer (25).

Continuous infusions of both propofoland remi-fentanyl are beneficial becauseboth are short acting. Propofol is safe inpatients with severe head injuries, is eas-ily titratable, and reduces ICP. In addi-tion, propofol decreases cerebral meta-bolic rate while having little effect onCPP reduction. Because of propofol’s ex-tremely short half-life, it is possible toarouse the patient in order to conduct athorough neurologic examination, andconsequently decrease both the numberof serial CT scans and the associated cost(25).


Kelly et al. (98) conducted a multi-center, double-blind trial in 42 intu-bated patients with head trauma inwhich continuous infusion of 2%propofol was compared with a regimenof morphine sulfate. Mean daily ICP andcerebral perfusion pressure were gener-ally similar between groups until thethird day of therapy, when ICP was sig-nificantly lower in the propofol groupcompared with the morphine group (p� .05). Patients treated with propofolrequired significantly less use of neuro-muscular blocking agents, benzodiaz-epines, pentobarbital, and cerebrospi-nal fluid drainage compared withpatients treated with morphine (p �.05). A favorable outcome, defined asgood recovery or moderate disability,was observed 6 months postinjury in52% of patients receiving propofol andin 47% receiving morphine, whereasthe mortality rates were 17% and 21%,respectively. The best outcomes wereachieved in patients receiving the high-est doses of propofol for the longestduration. The authors noted that, de-spite a higher incidence of poor prog-nostic indicators in the propofol group,propofol-based sedation, together withan ICP control regimen, is safe, accept-able, and is possibly a desirable alterna-tive to opioid-based sedation regimensin this patient population.

In another study (99), propofol wasevaluated in 10 patients with severehead injuries who were undergoing me-chanical ventilation. The rate of infu-sion of the drug was adjusted to main-tain the ICP at �10 mm Hg and CPP at60 mm Hg. Propofol was discontinuedafter 24 hrs. There were no significantdifferences in mean arterial pressure,but mean CPP tended to increase dur-ing the study. Overall, the quality ofsedation was determined to be good innine patients. In a study by Pearson etal. of hemodynamically stable headtrauma patients, propofol and mor-phine were compared for their effectson ICP and CPP. ICP was similar inboth treatment groups. In patientstreated with propofol, CPP increasedslightly over 48 hrs, whereas there wasa slight decrease in CPP in patientsreceiving morphine. The observed de-crease in ICP and increase in CPP withpropofol is consistent with the majorityof reports from other clinical trials andin the literature.

QUESTIONS AND ANSWERSFROM AGITATION ROUNDTABLEMEETING

Question

Dr. Cohen: It has been said that if youcannot measure it, you cannot manage it.An argument could be made that acuterespiratory distress syndrome manage-ment has improved a great deal becauseof the availability of better monitoringsystems such as pulse oximetry. Alongthis line of reasoning, what should we belooking for in the way of future monitor-ing tools to help fine-tune our approachto handling agitation in the ICU?

Answer

Dr. Gallagher: To date, the BIS moni-tor has been used to measure anestheticdepth in the operating room. In the ICU,BIS may only be useful in the paralyzedpatient—one that requires titration tosome sedation level. In nonparalyzed pa-tients, we tend to titrate to a clinicallevel—this can result in variable and con-tinually changing BIS levels—dependingon the degree of stimulation. There areothers in my group that have not yetfound it very useful, however; we stillrequire better, more objective monitor-ing of patient sedation and anxiety levels.

Another factor is an easily reproduc-ible sedation scoring system. For in-stance, with the Ramsey scale, the termi-nology goes back and forth between anexam and patient activity, and can beeasily misinterpreted.

We use a modified, modified Ramsayscale that everyone in our unit under-stands and interprets exactly the sameway. This has significantly improved se-dation titration and communication be-tween staff. Staff consensus regarding se-dation scale selection or adaptation isvery important.

Anne Pohlman: The key to all moni-toring devices or assessment tools is theability for all individuals using the de-vices to communicate the informationgained in a reliable and efficient manner.

Dr. Gallagher: The simpler the scale,the easier it is to reproduce. This makes itmuch more likely to be used and under-stood by everybody in the unit.

Anne Pohlman: One other piece ofequipment that is beginning to surface inthe ICU by way of our “Sleep” colleaguesis portable polysomnography equipment.As discussed earlier, the role of sleep in

the ICU is only beginning to be investi-gated in detail. Important to rememberwhen discussing new equipment or newtechnology, is our ability to incorporate itinto bedside practice in a manner thatassists the staff in caring for the patientto improve outcomes.

Question

Dr. Cohen: Based on the known evi-dence, what advice do you offer aboutroom lighting pattern, visiting hours, andtiming of nursing activities to prevent ortreat agitation?

Answer

Anne Pohlman: Control of environ-mental factors such as noise, lighting,room temperature, and around-the-clockstimulation from staff is clearly impor-tant in the treatment of agitated patients,but to date have not been studied orproven to change outcomes in acutely illpatients. Studies are underway to look atthe relationship of these variables in boththe chronic and acute critically ill patientin the ICU.

Dr. Papadakos: My personal opinion isthat sleep becomes extremely importantin the later stages of an illness. As anexample, when you have a patient that isdifficult to wean off the ventilator, a reg-ular sleep-wake cycle goes a long waytoward orienting the patient and facili-tates weaning.

Dr. Gallagher: The patient who issleep-deprived is probably much moredifficult to wean, but I do not think thisassociation has been well studied. We tryto keep patients who are weaning off theventilator comfortable because weaningrequires a large amount of work, and arested patient is a better candidate to beweaned.

Prof. Dasta: I think there is some im-mune response activity to sleep andmaybe wound healing relative to physio-logic sleep. So, if there are ways of mim-icking physiologic sleep, that would be agood attribute.

Question

Dr. Cohen: Anne, would you commenton how much literature there is to sup-port this information. What do you be-lieve to be profitable areas for future re-search into the topic of environmentalmanipulation and its impact on outcome?


Answer

Anne Pohlman: There is not muchliterature available with respect to envi-ronment changes in the ICU. The com-ponents of the environment that havebeen studied include excessive noise, ab-normal light/dark cycles, and frequentcare-related activities. These studies havetold us what many of us in the criticalcare world already know: it is noisy,bright, and patients do not get much un-interrupted time while in the ICU. Theeffectiveness of sleep-promoting strate-gies needs to be demonstrated, recogniz-ing the difficulty and complexity of doingthis type of study in the ICU. There are afew studies in the procedure areas ad-dressing interventions such as musictherapy, massage therapy, and therapeu-tic touch. The direct effect of these ma-neuvers on acutely ill patients remainsundetermined. Controlling the environ-ment in the ICU for noise, light, andtemperature is an ongoing challenge, asmany older ICUs do not have options formodifying temperature, light and noise,nor can bedside staff regulate them.Newer pumps, bedside monitors, andventilators allow bedside clinicians to setvolume and tone alarms to decreasenoise. Recent ICU room and unit designsallow for natural light from windows, andfor artificial lighting to be directed andcontrolled from wall dimmer switches.Portable unit-specific phones have re-cently been added to the ICU environ-ment; these phones are tied into the calllight system and the pager system set tovibrate rather than ring thus eliminatingthe need for unit intercoms.

Question

Dr. Cohen: Does anyone have any po-lices or rules and regulations governingenvironmental management for agitationin their ICUs?

Answer

Anne Pohlman: I have not seen anyhospital policies or rules regulating anyof these specific environmental concerns.However, in the 15 yrs that I have been anICU nurse, we have changed dramaticallythe policy regarding visiting hours andfamily involvement in care. Unit-specificenvironmental changes may include re-scheduling tasks that interrupt sleepsuch as baths in the middle of the night,4 a.m. daily chest x-rays, and scheduled

line changes during off hours. Poten-tially, these tasks could be “batched” intosingle time periods rather than occurringas continuous stimulation around theclock. Combining these tasks would de-pend of course on patient acuity, staffingpatterns, and medical staff availability forprocedures. With multidisciplinary “buyin,” it seems a study looking at theseissues may be possible.

Question

Dr. Cohen: Do any of the other panel-ists feel that study of environmental man-agement would be a worthwhile area ofinvestigation?

Answer

Dr. Papadakos: Yes, obviously, butmeasuring the impact of those variablesis going to be very, very difficult. Therehave been several studies in anesthesiatrying to put music headphones on thepatients during anesthesia and trying tomeasure whether or not that affects howthe patient feels afterward. But I think itmakes empirical sense that a very calm,soft environment is a lot better than anoisy, loud, bright environment. Youwould also have to look at the color onthe walls and the view outside of thewindow. My entire surgical ICU overlooksthe cemetery.

Anne Pohlman: Other ICU environ-mental issues such as encouraging familysupport and addressing psychosocialneeds of patients is imperative when deal-ing with agitated patients. In a recentlypublished paper by Hupcey [Hupcey JE:Feeling safe: The psychosocial needs ofICU patients. J Nurs Scholarsh 2000; 32:361–367], it was reported that the over-whelming need of ICU patients was to feelsafe. Family and friends, ICU staff, reli-gious beliefs, and feelings of knowing,regaining control, hoping, and trustingall influenced the perception of feelingsafe. Altering the ICU environment to fos-ter communication and address individ-ual patient/family needs during curativeinterventions or comfort-care strategiesis imperative.

Dr. Gallagher: Visitors are a two-edgedsword. We have fairly liberal visitinghours, but I do not think we pay enoughattention to who is visiting. Some visitorstend to help and others make thingsworse. This is a very difficult issue tograsp. It is a problem.

Question

Dr. Cohen: Translating knowledgeinto action is a serious concern in healthcare. What is our status in the area ofusing pharmacologic agents for agitationmanagement? Is there a problem, andwhat are its causes?

Answer

Prof. Dasta: In the Hansen-Flaschen etal. article in JAMA in 1991 [Hansen-Flaschen JH, Brazinsky S, Basile C, et al:Use of sedating drugs and neuromuscularblocking agents in patients requiring me-chanical ventilation for respiratory fail-ure. A national survey. JAMA 1991; 266:2870–2875], head nurses of pulmonaryICUs were asked what kinds of drugs wereused in their facilities. They reported thata wide variety of drugs were being em-ployed, and this awakened us to the prob-lem of polypharmacy in the agitated pa-tients. And I am not sure that the currentstate of affairs is that much better today,although I believe that guideline develop-ment with multidisciplinary input doesadd an evidence-based approach to whatwe do—if it is followed.

We surveyed this practice in our sur-gical ICU and published our results in1994. On average, our patients receivedtwo drugs. The range was zero to nine—one patient received nine different drugsfor agitation or pain. Overall we docu-mented 23 different drugs in more than200 patients.

We tend to throw things at patientswithout optimizing any single strategy,for instance adding a sedative while thepatient’s pain is not properly controlled.

With respect to our understanding ofthe metabolism and excretion of variousagents, we certainly know more aboutthem today than we did 10 yrs ago, butwe know about the kinetics and dynamicsin isolation, with monotherapy. What I donot think we fully understand is the dy-namics and kinetics of lorazepam, for ex-ample, in the patient who is also receiv-ing morphine and haloperidol, and alsodiphenhydramine. So, the complex phar-macology in the real world is poorly un-derstood.

Understanding what the various drugsdo and what they do not do remains anissue. It is not unusual to find a patientreceiving a neuromuscular blocker by in-fusion and not having a drug that hasamnestic properties on board. Or to see aPRN morphine order in a patient receiv-


ing a neuromuscular blocker. Such over-sights indicate a basic gap in understand-ing.

Anne Pohlman: The ability to changepractice based on research or evidence-based medicine continues to be an ongo-ing challenge. Using our own sedation“wake up” study as an example, when thestudy was completed and daily “wake up”assessments directed by the researchteam stopped—within a very short periodof time the recent practice change ofdaily “wake up” assessments stopped aswell. Fostering the practice changethrough a multidisciplinary approach inwhich there is “buy in” from all bedsidecaregivers was required. In our unit, se-dation “wake up” assessments are now apart of shift-to-shift report and dailyrounds.

Prof. Dasta: That is a really good point.Another issue is the long-term psycho-logical effects of sedatives. This is an areawe need to learn more about. In CriticalCare Medicine this past year, a study byNelson et al. (100) showed a correlationbetween the number of days of sedationand the development of depression aswell as post-traumatic stress disordersymptoms.

Question

Dr. Cohen: Withdrawing sedatives af-ter prolonged use frequently is an ardu-ous endeavor peppered by PRN sedativedoses and increasing drip rates. Becausedelirium is a likely occurrence in theawakening amnestic patient, what is therole for antipsychotic agents/haloperidolin combination with sedative agents? Andhow early should we be starting theseantipsychotic agents?

Answer

Dr. Abraham: Yes, and we touched onthe use of haloperidol earlier in this dis-cussion. This is actually a very compli-cated question. The sedative agents oftencause disorientation, particularly thebenzodiazepines, in critically ill patients,in the elderly patients, or in patients whohave multiple organ system dysfunction.Haloperidol is probably preferred in thissituation and should probably be insti-tuted at an earlier point. Ventilator-dependent patients, as their respiratorystatus is improving, are prone to havingsleep disorders and confusion, even post-extubation. Getting a handle on their de-

lirium at an early point with haloperidolis probably a good idea.

Dr. Papadakos: I think one of theother things that you have to consider iswithdrawal of opioids. As the associatedirector of a burn trauma unit, I can tellyou that we have a big problem trying towean people off long-term, high-dose opi-oids. We have had some success usingmethadone.

Dr. Abraham: Yes, and in the sameway I think patients who are on benzodi-azepines often develop tachyphylaxis, andso decreasing those agents contributes toincreased agitation as well.

Question

Dr. Cohen: What do you recommendfor sleep in the delirious patient forwhom you are trying to minimize the useof sedating agents?

Answer

Dr. Papadakos: We have done somework looking at increasing levels ofpropofol at night and then titrating it offin patients who are sleep deprived. Wehave used drugs such as diphenhydra-mine and some of the other commonlyprescribed sleep medications in patientswho are not intubated or on mechanicalventilation. We do use some of the sleepmedications that the elderly commonlytake, such as diphenhydramine, alprazo-lam, and shorter-acting benzodiazepines.Many elderly patients have trouble sleep-ing and are on drugs at home. We try toreplicate these regimens orally or via thefeeding tube. I do not know if other peo-ple on the panel use these agents in theirintensive care unit.

Prof. Dasta: We occasionally use di-phenhydramine, with caution in the el-derly because of the anticholinergic prop-erties that it might have. I know I take itwhen I am on an airplane and I want tosleep.

Question

Dr. Cohen: From a pathophysiologicalperspective, what do you believe are themost common errors made in the preven-tion or treatment of agitation?

Answer

Dr. Gallagher: I think there are two.First is failure to recognize pain. Second(especially in a training institution wheremultiple people are writing orders) every-

body has a favorite or preferred drug touse for the agitated patient. Usually thepatient ends up on a variety of agents. AsI stated earlier, once you get involvedwith multiple drugs, the situation be-comes confused and very difficult to sortout. So, I recommend simplicity in selec-tion and using only one agent at a time.

Question

Dr. Cohen: One of the principles oftotal quality management is that reduc-tion of variation leads to improvement inquality. In the past, regulatory standardshave forbidden the use of standardizedprotocol approaches to the use of re-straints. It sounds like the requirementsfor restraining agitated patients are nowgetting more onerous. Is this antiproto-col sentiment still the case, and how dowe deal with this counter-intuitive pol-icy?

Answer

Anne Pohlman: The regulatory stan-dards regarding the use of restraints havebeen revised again in 2001. Institution-specific policies and guidelines requiredto meet these revised standards can beonerous. The bottom line with the use ofrestraints remains unchanged; patientsafety is a priority and we, as healthcareproviders, need to optimize our treat-ment strategies to assure safety withoutmerely turning to restraints and tyingpatients down. So, in essence, as a bed-side practitioner, my goal—like the reg-ulatory standard—is patient safety. Theonerous part for us at the bedside is de-veloping guidelines that describe ourpractice of ensuring safety in the ICU. Forexample, (1 standardizing patient assess-ment tools; (2 instituting patient-specificinterventions that ensure patient safety(this may include the use of restraints ifnecessary, however other less restrictivemeasures should be initiated first); (3scheduling frequent reassessment prac-tices; and, of course, (4 documentationstrategies that confirm that this practiceis being carried out.

Dr. Cohen: It is my impression frompast reading of JCAHO literature thatthey expressly forbid the use of protocols.

Anne Pohlman: JCAHO forbids the useof protocols if it is a protocol that doesnot take individual patient needs into ac-count. For example, a protocol that statesthat all mechanically ventilated patientsrequire restraints while in the ICU is not


acceptable. However, a protocol statingthat patients in the ICU requiring me-chanical ventilation will be reassessedfrequently for level of consciousness andtolerance to the current therapies andinterventions is acceptable. Other itemsto consider in the protocol include sug-gested interventions to reorient the pa-tient and optimize safety with regard toinvasive catheters and interventions. Ifrestraints are required, they are appliedto ensure patient safety, not staff conve-nience.

Dr. Cohen: Can you use a protocolproviding that the standard is met withadequate documentation?

Anne Pohlman: Exactly. What they arereally looking for is an assessment ofwhat was going on, what you did, andwhat follow-up measures were taken.They want to ensure that we are not justrandomly putting patients in restraints.

Dr. Gallagher: We have a protocol thatworks fairly well. An intubated patient inthe ICU essentially meets the protocol.When the JCAHO came through the lasttime they were reasonably happy withthat approach.

Question

Dr. Cohen: About the risk of with-drawal from analgesics and sedatives, canyou comment on which agents, if any,present significant risk, and on the im-pact of duration and intensity of therapy?

Answer

Prof. Dasta: Opioid and benzodiaz-epine withdrawal after a week or more oftherapy presents particular problems.Maybe one way of addressing the problemis a progressive tapering or a systematictapering of therapy during the ICU stay.The daily awakening approach that Annehas used at her institution may also min-imize the risk of withdrawal.

Question

Dr. Cohen: Does anybody rotate orchange drug groups to try to avoid therisk of withdrawal?

Answer

Multiple Responses: No.Dr. Abraham: The important thing is

to try to judge when the patient needs tobe off the drug, and start some weaningprocess in advance of that—maybe sev-eral days, because the longer you go, the

longer, obviously, it is going to take. Butthat works much better than trying tochange drugs.

Anne Pohlman: The addition of theoral agents, for example, oral lorazepam,may help while the infusions are titratedoff slowly. We need to also remember thatmany of these patients may have beentaking medications like the benzodiaz-epines for anxiety before being hospital-ized as well.

Dr. Papadakos: I think I mentionedthat looking at the patient’s home medi-cations is important. As you pointed out,many of the elderly are on either psychi-atric medications or sleep medications athome, and trying to replicate those is key.

Question

Dr. Cohen: Would you briefly describeone or two examples (not necessarilyabout agitation) of approaches to reduc-ing the evidence-care gap and minimiz-ing variation that work in your criticalcare unit? By evidence-care gap, I meansomething that has been shown to workwell in the research setting, but its notbeing used well in practice.

Answer

Dr. Abraham: I think the best exampleof this is probably ventilator therapies foradult respiratory distress syndrome(ARDS). At the American Thoracic Soci-ety annual meeting, there was a fascinat-ing abstract from Gordon Rubenfeldlooking at utilization of low tidal volumeventilation at the University of Washing-ton (along with our institution, one often of the NIH ARDS network centers).Before the results showing a substantialbenefit were published in the New En-gland Journal of Medicine, about 5% ofthe patients were managed with low tidalvolume, since the New England Journalpaper appeared, about 8% are managedappropriately!

So, there is the issue about dissemi-nation of clinical management criteriaand utilization of these algorithms. Whatwe have done is institute protocols in themedical and surgical ICU for patientswith acute lung injury and ensured thatthe respiratory therapists are aware ofboth the protocols and which patients areat risk, and will actually follow up withthe radiologist to find out if the patientsdo, indeed, have acute lung injury. Thedirectors of the units have signed on tothis plan. The teams managing the pa-

tients have to make an active decision forthose patients not to be managed by theNIH protocol. We, like the University ofWashington and a lot of other people,were disappointed by voluntary assump-tion of these kinds of new clinical prac-tice patterns.

I believe that an institutional protocolis probably the best approach. As we dis-cussed with sedation, whenever one insti-tutes a protocol and standardizes therapy,there are multiple potential benefits.

Dr. Cohen: I think that 5% to 8%indicates just how big a problem thisreally is.

Dr. Abraham: Yes, it is tremendous; ina condition with a clear-cut 25% reduc-tion in mortality, and with publication ofa lead article in a major journal. Finally,there has been a lot of talking about it.Everybody says that they do manage theirpatients with low tidal volume ventila-tion, but the evidence says otherwise.

Dr. Papadakos: I think the availabilityof protocol order sheets that eliminatephysician variability is very important.That is what we’ve instituted at our insti-tution. The sedation protocol is actuallyan order check box in the patient’s chart.In teaching hospitals, this may be moredifficult because of the number of peopleinvolved.

We try to remove original thoughtfrom the process and do the same thingwith ventilator management. A low tidalvolume protocol has been used at theUniversity of Rochester for almost 12 yrs.It is on a standardized order sheet, sothere are not a lot of choices for theresidents or other staff members to make.

I wonder if supporting orders–youknow, protocolized orders—are probablymore important than having protocols.

Anne Pohlman: Holding all staff ac-countable to the practice change is im-portant. Making change a part of patientmanagement is optimal. For example, inthe case of low tidal volume ventilation,the respiratory therapist ensures compli-ance with ventilator settings during rou-tine ventilator checks. This way we havemultiple checks and balances within thesystem—physician orders, bedside moni-toring by nurses, and respiratory therapyvent checks—all making sure that theright strategy is in place.

Dr. Gallagher: I do not know that Inecessarily agree with protocols to theextent that boxes are “checked off,” par-ticularly in a training institution. Thisapproach can diminish the trainees’ abil-ities to think. When they get to a situa-


tion that does not match up, they canbecome lost and are unable to deal with itproperly. I do not think having the pro-tocol is wrong; the issue is whether youreview the patient’s course so that peoplecan learn also by their mistakes. It maybe a lot easier to run a protocol in theprivate setting, especially if just a fewindividuals are involved in overseeing theday-to-day management. In my experi-ence, to achieve multiple goals in a train-ing institution, dealing with protocols isa lot harder and takes a lot more energy.

Question

Dr. Cohen: Do you have any way ofproviding feedback for clinicians abouttheir compliance? As you said, everybodythinks they are doing a wonderful job. Doyou have any way of saying “your compli-ance with this protocol is 50%”; is thereany tracking?

Answer

Dr. Abraham: Yes, for example, withthe ventilated protocol, we do not do it onan individual basis, because people findthat a bit too threatening, but we do it ona divisional basis. Division members whoattend in the ICU, for example, will bemade aware of these statistics. In ourexperience, the drivers for these changesin practice are actually the fellows farmore than the faculty. The fellows andresidents are much more modifiable interms of their behavior patterns. And sowe also disseminate this information tothe house staff, both our own fellows andthe medicine house staff who are in theunit. The surgeons and collaborators dothe same thing.

Question

Dr. Cohen: A large number of peopleuse alternative therapies on a regular ba-sis in the outpatient setting. Can youcomment on the use of herbal medica-tions, acupuncture, melatonin, acupres-sure massage, etc., in the ICU?

Answer

Dr. Papadakos: Fortuitously, I was re-cently at the European meeting of anes-thesiology in Florence, and I did attend asmall satellite session on the use ofherbal medications. Many clinicians havea very low understanding of the pharma-cologic effects of many herbal medica-tions.

We now know about the drug–druginteraction of some commonly usedherbal medications in our society. I thinkit is important for us, before we startusing these drugs, to develop an under-standing of how they work. Educatinghealthcare providers in how these drugswork is a necessary first step.

Other alternative therapies (acupunc-ture and acupressure), are commonlyused in the critical care settings in hos-pitals in the Orient. I have rounded on anintensive care unit in China, and I cantell you that acupuncture is used com-monly and successfully in patients formanagement of pain and reduction ofanxiety. But again, the same level of ex-pertise does not exist in Western medi-cine.

Overall, I think more and more peopleare using herbal medications and herbalteas, and I think there is a growing inter-est to educate physicians in the interac-tion of herbal therapies and traditionalpharmacologic agents.

REFERENCES

1. Covington H: Use of propofol for sedation inthe ICU. Crit Care Nurse 1998; 18:34–39

2. Riker RR, Picard JT, Fraser GL: Prospectiveevaluation of the Sedation-Agitation Scalefor adult critically ill patients. Crit CareMed 1999; 27:1325–1329

3. Deleted in proof4. Wise MG, Cassem NH: Behavioral distur-

bances. In: Critical Care. Third Edition.Civetta JM, Taylor RW, Kirby RR (Eds).Philadelphia, Lippincott-Raven, 1997, pp2017–2029

5. Crippen D, Ermakov S: Stress, agitation,and brain failure in critical care medicine.Crit Care Nurs Q 1992; 15:52–74

6. Harvey M: Managing agitation in criticallyill patients. Am J Crit Care 1996; 5:7–15

7. Haskell RM, Frankel HL, Rotondo MF. Agi-tation. AACN Clin Issues 1997; 3:335–350

8. Sessler CN, Grap MJ, Brophy GM: Multidis-ciplinary management of sedation and an-algesia in critical care. Semin Respir CritCare Med 2001; 22:211–225

9. Levine RL: Pharmacology of intravenoussedatives and opioids in critically ill pa-tients. Crit Care Clin 1994; 10:709–731

10. Jenkins DH: Substance abuse and with-drawal in the intensive care unit: Contem-porary issues. Surg Clin North Am 2000;80:1033–1053

11. Hassan E, Fontaine DK, Nearman HS: Ther-apeutic considerations in the managementof agitated or delirious critically ill patients.Pharmacotherapy 1998; 18:113–129

12. Crippen DW: The role of sedation in the ICUpatient with pain and agitation. Crit CareClin 1990; 6:369–392

13. DeJonghe B, Cook D, Appere-De-Veechi C,et al: Using and understanding sedationscoring systems: A systematic review. Inten-sive Care Med 2000; 26:275–285

14. Brook AD, Ahrens TS, Schaiff R, et al: Effectof a nursing-implemented sedation protocolon the duration of mechanical ventilation.Crit Care Med 1999; 27:2609–2615

15. Hansen-Flashen J, Cowen J, Palomano RC:Beyond the Ramsay scale: Need for a vali-dated measure of sedating drug efficacy inthe intensive care unit. Crit Care Med 1994;22:732–733

16. Devlin JW, Boleski G, Mlynarek M, et al:Motor activity assessment scale: A valid andreliable sedation scale for use with mechan-ically ventilated patients in an adult surgi-cal intensive care unit. Crit Care Med 1999;27:1271–1275

17. Ely EW, Margolin R, Francis J, et al: Eval-uation of delirium in critically ill patients:Validation of the Confusion AssessmentMethod for the Intensive Care Unit (CAM-ICU). Crit Care Med 2001; 29:1370–1379

18. Riker RR, Fraser GL: Monitoring sedation,agitation, analgesia, neuromuscular block-ade, and delirium in adult ICU patients.Semin Respir Crit Care Med 2001; 22:189–198

19. Gan TJ, Glass PS, Windsor A, et al: Bispec-tral index monitoring allows faster emer-gence and improved recovery from propo-fol, alfentanil, and nitrous oxide anesthesia.Anesthesiology 1997; 87:808–815

20. Kearse LA Jr, Rosow C, Zaslavsky A, et al:Bispectral analysis of the electroencephalo-gram predicts conscious processing of in-formation during propofol sedation andhypnosis. Anesthesiology 1998; 88:25–34

21. Weinart CR, Chlan L, Gross C: Sedatingcritically ill patients: factors affectingnurses’ delivery of sedative therapy. Am JCrit Care 2001; 10:156–167

22. Cooper AB, Gabor JY, Hanly PJ: Sleep in thecritically ill patient. Semin Respir Crit CareMed 2001; 22:153–163

23. Kress JP, Pohlman AS, O’Connor MF, et al:Daily interruption of sedative infusions incritically ill patients undergoing mechani-cal ventilation. N Engl J Med 2000; 342:1471–1477

24. Joint Commission on Accreditation ofHealthcare Organizations (JCAHO): Com-prehensive Accreditation Manual for LongTerm Care. Standards and Intents for Seda-tion and Anesthesia Care. Revisions to An-esthesia Care Standards, effective January 1,2001

25. Hooper VD, George-Gay B: Sedation in thecritically ill patient. Crit Care Nurs ClinNorth Am 1997; 9:395–410

26. Mazzeo AJ: Sedation for the mechanicallyventilated patient. Crit Care Clin 1995; 11:937–955

27. Young C, Knudsen N, Hilton A, et al: Seda-tion in the intensive care unit. Crit CareMed 2000; 28:854–866

28. Mascia MF, Koch M, Medicis JJ, et al: Phar-


macoeconomic impact of rational useguidelines on the provision of analgesia,sedation, and neuromuscular blockade incritical care. Crit Care Med 2000; 28:2300–2306

29. Doyle DJ: Propofol: A synopsis of its phar-macology and clinical uses in the OR andICU. Available at: http://doyle.ibme.utoron-to.ca/propofol/. Accessed May 23, 2001

30. Ostermann ME, Keenan SP, Seiferling RA,et al: Sedation in the intensive care unit: Asystematic review. JAMA 2000; 283:1451–1459

31. Spencer L, Willatts S: Sedation. In: Text-book of Intensive Care. Goldhill D, Withing-ton S (Eds). London, Chapman & Hall,1997, pp 95–101

32. American College of Critical Care Medicine,American Society of Health-System Phar-macists, American College of Chest Physi-cians: Clinical practice guidelines for thesustained use of sedatives and analgesics inthe critically ill adult. Crit Care Med 2002;30:119–141

33. Mascia MF, Koch M, Medicis JJ, et al: Phar-macoeconomic impact of rational useguidelines on the provision of analgesia,sedation, and neuromuscular blockade incritical care. Crit Care Med 2000; 28:2300–2306

34. Devlin JW, Holbrook AM, Fuller HD: Theeffect of ICU sedation guidelines and phar-macist interventions on clinical outcomesand drug cost. Ann Pharmacother 1997;31:689–695

35. Jacobi J, Farrington EA: Supportive care ofthe critically ill patient. In: Pharmacother-apy Self-Assessment Program (PSAP). ThirdEdition. Kansas City, MO: The AmericanCollege of Clinical Pharmacy; 1999, pp129–151

36. Wagner BKJ, O’Hara DA: Pharmacokineticsand pharmacodynamics of sedatives and an-algesics in the treatment of agitated criti-cally ill patients. Clin Pharmacokinet 1997;33:425–453

37. Mirski MA, Muffelman B, Ulatowski JA, et al:Sedation for the critically ill neurologic pa-tient. Crit Care Med 1995; 23:2038–2053

38. Shapiro BA, Warren J, Egol AB, et al: Prac-tice parameters for intravenous analgesiaand sedation for adult patients in the inten-sive care unit: An executive summary. CritCare Med 1995; 23:1596–1600

39. Nasraway SA: Use of sedative medications inthe intensive care unit. Semin Respir CritCare Med 2001; 22:165–174

40. Watling SM, Dasta JF, Seidl EC: Sedatives,analgesics and paralytics in the ICU. AnnPharmacother 1997; 31:148–153

41. Shafer A: Complications of sedation withmidazolam in the intensive care unit and acomparison with other sedative regimens.Crit Care Med 1998; 26:947–956

42. Cawley MJ: Short-term lorazepam infusionand concern for propylene glycol toxicity:Case report and review. Pharmacotherapy2001; 21:1140–1144

43. Lown KS, Thummel KE, Benedict PE, et al:The erythromycin breath test predicts theclearance of midazolam. Clin PharmacolTher 1995; 57:16–24

44. Pohlman AS, Simpson KP, Hall JB: Contin-uous intravenous infusions of lorazepamversus midazolam for sedation during me-chanical ventilatory support: A prospective,randomized study. Crit Care Med 1994; 22:1241–1247

45. Mirenda J, Broyles J: Propofol as used forsedation in the ICU. Chest 1995; 108:539–548

46. Cremer OL, Moons KGM, Bouman EAC, etal: Long-term propofol infusion and cardiacfailure in adult head-injured patients. Lan-cet 2001; 357:117–118

47. Aitkenhead AR, Willatts SM, Park GR, et al:Comparison of propofol and midazolam forsedation in critically ill patients. Lancet1989; 2:704–709

48. Carrasco G, Cabre L, Sobrepere G, et al:Synergistic sedation with propofol and mi-dazolam in intensive care patients after cor-onary artery bypass grafting. Crit Care Med1998; 26:844–851

49. Chamorro C, de Latorre FJ, Montero A, etal: Comparative study of propofol versusmidazolam in the sedation of critically illpatients: Results of a prospective, random-ized, multicenter trial. Crit Care Med 1996;24:932–939

50. Grounds RM, Lalor JM, Lumley J, et al:Propofol infusion for sedation in the inten-sive care unit: preliminary report. BMJ1987; 294:397–400

51. Hall RI, Sandham D, Cardinal P, et al:Propofol vs midazolam for ICU sedation: ACanadian multicenter randomized trial.Chest 2001; 119:1151–1159

52. Higgins TL, Yared J-P, Estafanous FG, et al:Propofol versus midazolam for intensivecare unit sedation after coronary bypassgrafting. Crit Care Med 1994; 22:1415–1423

53. Kress JP, O’Conner MF, Pohlman AS, et al:Sedation of critically ill patients during me-chanical ventilation: A comparison ofpropofol and midazolam. Am J Respir CritCare Med 1996; 153:1012–1018

54. McMurray TJ, Collier PS, Carson IW, et al:Propofol sedation after open heart surgery:A clinical and pharmacokinetic study. An-aesthesia 1990; 45:322–326

55. Roekaerts PMHJ, Huygen FJPM, de LangeS: Infusion of propofol versus midazolamfor sedation in the intensive care unit fol-lowing coronary artery surgery. J Cardio-thorac Vasc Anesth 1993; 7:142–147

56. Ronan KP, Gallagher TJ, George B, et al:Comparison of propofol and midazolam forsedation in intensive care unit patients. CritCare Med 1995; 23:286–293

57. Sanchez-Izquierdo-Riera JA, Caballero-Cubedo RE, Perez-Vela JL, et al: Propofolversus midazolam: Safety and efficacy forsedating the severe trauma patient. AnesthAnalg 1998; 86:1219–1224

58. Sandiumenge Camps A, Sanchez-Izquierdo

Riera JA, Vasquez DT, et al: Midazolam and2% propofol in long-term sedation of trau-matized, critically ill patients: Efficacy andsafety comparison. Crit Care Med 2000; 28:3612–3619

59. Barrientos-Vega R, Sanchez-Soria MM, Mo-rales-Garcia C, et al: Prolonged sedation ofcritically ill patients with midazolam orpropofol: Impact on weaning and costs. CritCare Med 1997; 25:33–44

60. Barrientos-Vega R, Sanchez-Soria M, Mo-rales-Garcia C, et al: Pharmacoeconomic as-sessment of propofol 2% used for prolongedsedation. Crit Care Med 2001; 29:317–322

61. Carrasco G, Molina R, Costa J, et al: Propo-fol vs midazolam in short-, medium-, andlong-term sedation of critically ill patients:A cost-benefit analysis. Chest 1993; 103:557–564

62. Deleted in proof63. Ewart MC, Yau KW, Morgan M: 2% propofol

for sedation in the intensive care unit: Afeasibility study. Anesthesia 1992; 47:146–148

64. McLeod G, Dick J, Wallis C, et al: Propofol2% in critically ill patients: Effect on lipids.Crit Care Med 1997; 25:1976–1981

65. Bennett SN, McNeil MM, Bland LA, et al:Postoperative infections traced to contami-nation of an intravenous anesthetic, propo-fol. N Engl J Med 1995; 333:147–154

66. Thompson KA, Goodale DB: The recent de-velopment of propofol (DIPRIVAN®). Inten-sive Care Med 2000; 26:S400–S404

67. Herr DL, Kelly K, Hall JB, et al: Safety andefficacy of propofol with EDTA when usedfor sedation of surgical intensive care unitpatients. Intensive Care Med 2000; 26:S452–S462

68. Higgins TL, Murray M, Kett DH, et al: Traceelement homeostasis during continuous se-dation with propofol containing EDTA ver-sus other sedatives in critically ill patients.Intensive Care Med 2000; 26:S413–S421

69. Redhead HM, Jones CB, Bobey DG: Pharma-ceutical and antimicrobial differences be-tween propofol emulsion products [letter].Am J Health Syst Pharm 2000; 57:1174–1175

70. Nejman AM: Sedation and paralysis. In:Critical Care. Third Edition. Civetta JM,Taylor RW, Kirby RR (Eds). Philadelphia,Lippincott-Raven, 1997, pp 821–835

71. Carroll K, Magruder K: The role of analge-sics and sedatives in the management ofpain and agitation during weaning frommechanical ventilation. Crit Care Nurs Q1993; 15:68–77

72. Lund N, Papadakos PJ: Barbiturates, neuro-leptics, and propofol for sedation. Crit CareClin 1995; 11:875–886

73. Gertler R, Cleighton Brown H, Mitchell DH,et al: Dexmedetomidine: A novel sedative-analgesic agent. BUMC Proc 2001; 14:13–21

74. Bhana N, Goa KL, McClellan KJ: Dexme-detomidine. Drugs 2000; 59:263–268

75. Silbert BS, Santamaria JD, O’Brien JL, andThe Fast Track Cardiac Care Team: Early


extubation following coronary artery bypasssurgery: A prospective randomized con-trolled trial. Chest 1998; 113:1481–1488

76. Roekaerts PMHJ, Huygen FJPM, de LangeS: Infusion of propofol versus midazolamfor sedation in the intensive care unit fol-lowing coronary artery surgery. J Cardio-thorac Vasc Anesth 1993; 7:142–147

77. Cheng DCH, Karski J, Peniston C, et al:Morbidity outcome in early versus conven-tional tracheal extubation after coronary ar-tery bypass grafting: A prospective random-ized controlled trial. J Thorac CardiovascSurg 1996; 112:755–764

78. Cheng DCH, Karski J, Peniston C, et al:Early tracheal extubation after coronary ar-tery bypass graft surgery reduces costs andimproves resource use: A prospective, ran-domized, controlled trial. Anesthesiology1996; 85:1300–1310

79. Roberts BL: Managing delirium in adult in-tensive care patients. Crit Care Nurse 2001;21:48–54

80. Fish DN: Treatment of delirium in the crit-ically ill patient. Clin Pharm 1991; 10:456–66

81. Justic M: Does “ICU psychosis” really exist?Crit Care Nurse 2000; 20:28–37

82. Yagan MB, White D, Stabb J: Sedation of themechanically ventilated patient. Crit CareNurs Q 2000; 22:90–100

83. Hess DR, Kacmarek RM: Pharmacologictherapy: Sedatives, analgesics, and paralyticagents. In: Essentials of Mechanical Venti-lation. Hess DR, Kacmarek RM (Eds). NewYork, McGraw Hill, 1996, pp 207–214

84. Valente JF, Anderson GL, Branson RD, et al:

Disadvantages of prolonged propofol seda-tion in the critical care unit. Crit Care Med1994; 22:710–712

85. Dattani D, Wong G: Intensive care unitsedation: Agents used to sedate agitatedICU patients. Available at: http://www.pharmacyconnects.com/content/;hppractice/1996/01–96/f04_features.html.Accessed August 4, 2001

86. Walder B, Elia N, Henzi I, et al: A lack ofevidence of superiority of propofol versusmidazolam for sedation in mechanicallyventilated critically ill patients: A qualita-tive and quantitative systematic review.Anesth Analg 2001; 92:975–983

87. McCowen C, Marik P: Refractory deliriumtremens treated with propofol: A case series.Crit Care Med 2000 6:1781–1784

88. Coomes TR, Smith SW: Successful use ofpropofol in refractory delirium tremens.Ann Emerg Med 1997; 30:825–828

89. Durbin CG Jr: Sedation of the agitated, crit-ically ill patient without an artificial airway.Crit Care Clin 1995; 11:913–936

90. Brody SL, Slovis CM, Wrenn KD: Cocaine-related medical problems: Consecutive se-ries of 233 patients. Am J Med 1990; 88:325–331

91. Padula C, Willey C: Tobacco withdrawal inCCU patients. Dimens Crit Care Nurs 1993;12:305–312

92. Conti G, Dell’Utri D, Vilardi V, et al: Propo-fol induces bronchodilation in mechanicallyventilated chronic obstructive pulmonarydisease (COPD) patients. Acta AnaethesiolScand 1993; 37:105–109

93. Zaloga GP, Todd N, Levit P, et al: Propofol-

induced bronchodilation in patients withstatus asthmaticus. Internet J Emerg Inten-sive Care Med 2001; 5:[http://www.icaap.org/iuicode?895.1.5

94. Faber-Langendoen K, Lanken PN, for theACP–ASIM End-of-Life Care ConsensusPanel: Dying patients in the intensive careunit: Forgoing treatment, maintainingcare. Ann Intern Med 2000; 133:886–893

95. Wilson WC, Smedira NG, Fink C, et al:Ordering and administration of sedativesand analgesics during the withholding andwithdrawal of life support from critically illpatients. JAMA 1992; 267:949–953

96. Hall RI, Graeme M, Rocker GM: End-of-lifecare in the ICU: Treatments provided whenlife support was or was not withdrawn.Chest 2000; 118:1424–1430

97. Beilin Y, Bodian CA, Mukherjee T, et al: Theuse of propofol, nitrous oxide, or isofluranedoes not affect the reproductive success ratefollowing gamete intrafallopian transfer(GIFT): A multicenter pilot trial/survey.Anesthesiology 1999; 90:36–41

98. Kelly DF, Goodale DB: Williams J, et al:Propofol in the treatment of moderate andsevere head injury: A randomized, prospec-tive double-blinded pilot trial. J Neurosurg1999; 90:1042–1052

99. Smith I, White PF, Nathanson M, et al:Propofol: an update on its clinical use.Anesthesiology 1994; 81:1005–1043

100. Nelson BJ, Weinert CR, Bury CL, et al: In-tensive care unit drug use and subsequentquality of life in acute lung injury patients.Crit Care Med 2000; 28:3626–3630


Management of the agitated intensive care unit patientedmedia.emory.edu/GStaton/Management of the...May 23, 2001 · scales, daily awakening, and other strat-egies take on more of

Documents