Guia De Sedacion

Clinical practice guidelines for the sustained use of sedatives andanalgesics in the critically ill adult

Judith Jacobi, PharmD, FCCM, BCPS; Gilles L. Fraser, PharmD, FCCM; Douglas B. Coursin, MD;Richard R. Riker, MD; Dorrie Fontaine, RN, DNSc, FAAN; Eric T. Wittbrodt, PharmD;Donald B. Chalfin, MD, MS, FCCM; Michael F. Masica, MD, MPH; H. Scott Bjerke, MD;William M. Coplin, MD; David W. Crippen, MD, FCCM; Barry D. Fuchs, MD; Ruth M. Kelleher, RN;Paul E. Marik, MDBCh, FCCM; Stanley A. Nasraway, Jr, MD, FCCM; Michael J. Murray, MD, PhD, FCCM;William T. Peruzzi, MD, FCCM; Philip D. Lumb, MB, BS, FCCM. Developed through the Task Force of theAmerican College of Critical Care Medicine (ACCM) of the Society of Critical Care Medicine (SCCM), incollaboration with the American Society of Health-System Pharmacists (ASHP), and in alliance with theAmerican College of Chest Physicians; and approved by the Board of Regents of ACCM and the Council ofSCCM and the ASHP Board of Directors

Maintaining an optimal levelof comfort and safety forcritically ill patients is auniversal goal for critical

care practitioners. The American Collegeof Critical Care Medicine (ACCM) of theSociety of Critical Care Medicine’s(SCCM’s) practice parameters for the op-timal use of sedatives and analgesics waspublished in 1995 and recommended atiered approach to the use of sedativesand analgesics, largely on the basis of ex-pert opinion (1). These clinical practiceguidelines replace the previously publishedparameters and include an evaluation ofthe literature published since 1994 com-paring the use of these agents. The readershould refer to the accompanying introduc-

tion for a description of the methodologyused to develop these guidelines (2).

This document is limited to a discus-sion of prolonged sedation and analgesia.Consistent with the previous practiceguidelines, this document pertains to pa-tients older than 12 years. The majorityof the discussion focuses on the care ofpatients during mechanical ventilation. Adiscussion of regional techniques is notincluded. Appendix A summarizes therecommendations made herein.

ANALGESIA

In these guidelines, “analgesia” is de-fined as the blunting or absence of sensa-tion of pain or noxious stimuli. Intensivecare unit (ICU) patients commonly havepain and physical discomfort from obviousfactors, such as preexisting diseases, inva-sive procedures, or trauma. Patient painand discomfort can also be caused by mon-itoring and therapeutic devices (such ascatheters, drains, noninvasive ventilatingdevices, and endotracheal tubes), routinenursing care (such as airway suctioning,physical therapy, dressing changes, and pa-tient mobilization), and prolonged immo-bility (3, 4). Unrelieved pain may contributeto inadequate sleep, possibly causing ex-haustion and disorientation. Agitation in anICU patient may result from inadequatepain relief. Unrelieved pain evokes a stressresponse characterized by tachycardia, in-creased myocardial oxygen consumption,hypercoagulability, immunosuppression,and persistent catabolism (5, 6). The com-

bined use of analgesics and sedatives mayameliorate the stress response in criticallyill patients (7, 8). Pain may also contributeto pulmonary dysfunction through local-ized guarding of muscles around the area ofpain and a generalized muscle rigidity orspasm that restricts movement of the chestwall and diaphragm (9). Effective analgesiamay diminish pulmonary complications inpostoperative patients (10).

Some patients recall unrelieved painwhen interviewed about their ICU stays (3,11, 12). The perception of pain can be in-fluenced by several factors, such as the ex-pectation of pain, prior pain experiences, apatient’s emotional state, and the cognitiveprocesses of the patient (11). Patientsshould be educated about the potential forpain and instructed to communicate theirneeds in an appropriate manner (such asusing an assessment tool or other commu-nication techniques). The goals of therapyshould also be communicated to the pa-tient and family. In many cases, pain will bemanaged but not completely eliminated.Fear of potent analgesics and misconcep-tions about pain and analgesics should beaddressed. Similarly, practitioner biasagainst the adequate use of opioids or mis-placed fears of adverse effects or addictionmay produce inadequate prescribing or ad-ministration (13, 14). Educating practitio-ners and assessing the quality of a painmanagement program may improve anal-gesia therapy, but such programs have notbeen universally successful (4, 15). The im-portance of appropriate pain managementprograms has been reinforced by the Joint

The American College of Critical Care Medicine(ACCM), which honors individuals for their achieve-ments and contributions to multidisciplinary criticalcare medicine, is the consultative body of the Societyof Critical Care Medicine (SCCM) that possesses rec-ognized expertise in the practice of critical care. TheCollege has developed administrative guidelines andclinical practice parameters for the critical care prac-titioner. New guidelines and practice parameters arecontinually developed, and current ones are system-atically reviewed and revised.

Special thanks to E. Wesley Ely, MD, for his con-tribution to the section on delirium.

Address correspondence to Society of Critical CareMedicine, 701 Lee Street, Suite 200, Des Plaines, IL60016. Available at www.sccm.org

Key Words: analgesia; sedation; evidence-basedmedicine; fentanyl; hydromorphone; morphine; loraz-epam; midazolam; propofol; haloperidol; guidelines

Copyright © 2002 by the Society of Critical CareMedicine and the American Society of Health-SystemPharmacists, Inc.

119Crit Care Med 2002 Vol. 30, No. 1

Commission on Accreditation of Health-care Organizations’s (JCAHO’s) establish-ment of standards on pain assessment andmanagement.

Recommendation: All critically ill pa-tients have the right to adequate an-algesia and management of their pain.(Grade of recommendation � C)

Pain Assessment

There is a limited amount of literaturethat directly addresses pain assessment inthe critical care unit. The articles re-viewed in this report include descriptionsof pain assessment tools used for criti-cally ill patients, even if these tools werenot validated in this population. Studiesof pain in the critically ill indicate theimportance of systematic and consistentassessment and documentation (16). Themost reliable and valid indicator of pain isthe patient’s self-report (17). The loca-tion, characteristics, aggravating and al-leviating factors, and intensity of painshould be evaluated. Assessment of painintensity may be performed with unidi-mensional tools, such as a verbal ratingscale (VRS), visual analogue scale (VAS),and numeric rating scale (NRS). VAScomprises a 10-cm horizontal line withdescriptive phrases at either end, from“no pain” to “severe pain” or “worst painever.” Variations include vertical divi-sions or numeric markings. VAS is reli-able and valid for many patient popula-tions (18). Though not specifically testedin the ICU, VAS is frequently used there(19–22). Elderly patients may have diffi-culty with VAS (20). NRS is a zero to tenpoint scale and patients choose a numberthat describes the pain, with ten repre-senting the worst pain. NRS is also valid,correlates with VAS, and has been used toassess pain in cardiac surgical patients(21). Because patients can complete theNRS by writing or speaking, and becauseit is applicable to patients in many agegroups, NRS may be preferable to VAS incritically ill patients.

Multidimensional tools, such as theMcGill Pain Questionnaire (MPQ) and theWisconsin Brief Pain Questionnaire(BPQ), measure pain intensity and thesensory, affective, and behavioral compo-nents of that pain but take longer toadminister and may not be practical forthe ICU environment (18, 22). The MPQand BPQ are reliable and valid tools buthave not been tested or used in the ICU.

Although the most reliable indicatorof pain intensity is what the patient re-

ports, critically ill patients are often un-able to communicate their level of pain ifsedated, anesthetized, or receiving neu-romuscular blockade. Neither the VASnor the NRS will resolve this problem asthey rely on the patient’s ability to com-municate with the care provider. Behav-ioral-physiological scales may be usefulin assessing pain in these patients. Mod-erate agreement was found between theVAS and the observer-reported Facesscale for all observations, but less agree-ment was noted as the pain intensity in-creased (19). The verbal descriptive scale(VDS) used in another trial showed mod-erate correlation (r � 0.60) with a behav-ioral pain scale in assessing pain in post-anesthesia patients (23). A behavioral-physiological scale was compared with anNRS and a moderate-to-strong correla-tion was observed between the scales(24). The behavioral-physiological scalealso assessed pain-related behaviors(movement, facial expression, and pos-turing) and physiological indicators(heart rate, blood pressure, and respira-tory rate). However, such nonspecific pa-rameters might be misinterpreted or af-fected by observer bias, leading to anunderestimation of the degree of pain ex-perienced by the patient (12, 24–27).

Family members or other surrogateshave been evaluated for their ability toassess the amount of pain experienced bynoncommunicative ICU patients. Whilesurrogates could estimate the presence orabsence of pain in 73.5% of patients, theyless accurately described the degree ofpain (53%) (28).

The most appropriate pain assessmenttool will depend on the patient involved,his/her ability to communicate, and thecaregiver’s skill in interpreting pain be-haviors or physiological indicators.

Recommendations: Pain assessmentand response to therapy should be per-formed regularly by using a scale ap-propriate to the patient populationand systematically documented.(Grade of recommendation � C)

The level of pain reported by the pa-tient must be considered the currentstandard for assessment of pain andresponse to analgesia whenever possi-ble. Use of the NRS is recommended toassess pain. (Grade of recommenda-tion � B)

Patients who cannot communicateshould be assessed through subjectiveobservation of pain-related behaviors(movement, facial expression, and pos-

Sedation and AnalgesiaTask Force

Chair, Sedation and Analgesia Task ForceJudith Jacobi, Pharm.D., FCCM, BCPSTask Force ChairCritical Care Pharmacy SpecialistMethodist Hospital, Clarian Health PartnersIndianapolis, INStanley A. Nasraway, Jr, MD, FCCMExecutive Director of Task ForceDirector, SICUAssociate Professor of Surgery, Medicine, and AnesthesiaTufts New England Medical CenterBoston, MAMembersH. Scott Bjerke, MDMedical Director of TraumaMethodist HospitalIndianapolis, INDonald B. Chalfin, MD, MS, FCCMDirector, Division of Research and Attending IntensivistDepartment of Emergency MedicineMaimonides Medical CenterAssociate Professor of Clinical Epidemiology andSocial MedicineAlbert Einstein College of MedicineBrooklyn, NYWilliam M. Coplin, MDAssociate Professor, Departments of Neurology &Neurological SurgeryWayne State UniversityChief, Neurology; Medical Director Neurotrauma andCritical CareDetroit Receiving HospitalDetroit, MIDouglas B. Coursin, MDProfessor of Anesthesiology and Internal MedicineAssociate Director of the Trauma and Life Support CenterUniversity of WisconsinMadison, WIDavid W. Crippen, MD, FCCMAssociate DirectorDepartments of Emergency and Critical Care MedicineSt. Francis Medical CenterPittsburgh, PADorrie Fontaine, RN, DNSc, FAANAssociate ProfessorAssociate Dean for Undergraduate StudiesGeorgetown UniversitySchool of Nursing and Health StudiesWashington, DCGilles L. Fraser, PharmD, FCCMDepartment of Critical Care MedicineMaine Medical CenterPortland, MEBarry D. Fuchs, MDMedical Director, MICUHospital of the University of Pennsylvania, UPHSPhiladelphia, PARuth M. Kelleher, RNDepartment of NursingNew England Medical CenterBoston, MAPhilip D. Lumb, M.B., B.S., FCCMProfessor and Chairman, Department of AnesthesiologyKeck School of Medicine of USCLos Angeles, CAPaul E. Marik, MDBCh, FCCMDepartment of Critical Care MedicineUniversity of Pittsburgh Medical SchoolPittsburgh, PAMichael F. Mascia, MD, MPHMedical Director, Outpatient Surgical ServicesDepartment of AnesthesiologyTulane University School of MedicineNew Orleans, LAMichael J. Murray, MD, PhD, FCCMDean, Mayo School of Health SciencesProfessor of Anesthesiology, Mayo Medical SchoolChair, Department of AnesthesiologyMayo ClinicJacksonville, FLWillam T. Peruzzi, MD, FCCMAssociate Professor of Anesthesiology,Northwestern University School of MedicineChief, Section of Critical Care Medicine,Northwestern Memorial HospitalChicago, ILRichard R. Riker, MD, Assistant ChiefDepartment of Critical Care MedicineMaine Medical CenterPortland, MEEric T. Wittbrodt, PharmDAssociate Professor of Clinical PharmacyPhiladelphia College of PharmacyUniversity of the Sciences in PhiladelphiaPhiladelphia, PA

120 Crit Care Med 2002 Vol. 30, No. 1

turing) and physiological indicators(heart rate, blood pressure, and respi-ratory rate) and the change in theseparameters following analgesic ther-apy. (Grade of recommendation � B)

Analgesia Therapy

Nonpharmacologic interventions in-cluding attention to proper positioning ofpatients, stabilization of fractures, andelimination of irritating physical stimu-lation (e.g., proper positioning of ventila-tor tubing to avoid traction on the endo-tracheal tube) are important to maintainpatient comfort. Application of heat or coldtherapy may be useful. Other nonpharma-cologic techniques to promote patientcomfort are discussed later in this docu-ment.

Pharmacologic therapies include opi-oids, nonsteroidal anti-inflammatory drugs(NSAIDs), and acetaminophen. Opioids me-diate analgesia by interacting with a varietyof central and peripheral opioid receptors.The opioids currently available have activityat a variety of these receptors, although the�- and �-receptors are most important foranalgesia. Interaction at other receptorsmay contribute to adverse effects. The an-algesic agents most commonly used in ICUpatients (fentanyl, morphine, and hydro-morphone) are addressed later (29). Al-though alfentanil has previously been re-ported as an analgesic with sedative effects,it will not be extensively discussed becauseit is not commonly used in North America(29).

Comparative trials of opioids have notbeen performed in critically ill patients.The selection of an agent depends on itspharmacology and potential for adverseeffects. The characteristics of commonlyused opioids and nonopioids are reviewedin Table 1 (30–32). Desirable attributes ofan opioid include rapid onset, ease oftitration, lack of accumulation of the par-ent drug or its metabolites, and low cost.Fentanyl has the most rapid onset andshortest duration, but repeated dosingmay cause accumulation and prolongedeffects. Morphine has a longer duration ofaction, so intermittent doses may begiven. However, hypotension may resultfrom vasodilation and an active metabo-lite may cause prolonged sedation in thepresence of renal insufficiency. Hydro-morphone’s duration of action is similarto morphine’s, but hydromorphone lacksa clinically significant active metaboliteor histamine release. Meperidine has anactive metabolite that causes neuroexci-

tation (apprehension, tremors, delirium,and seizures) and may interact with an-tidepressants (contraindicated withmonoamine oxidase inhibitors and bestavoided with selective serotonin-reuptakeinhibitors), so it is not recommended forrepetitive use (17, 33, 34). Because co-deine lacks analgesic potency, it is notuseful for most patients. Remifentanilhas not been widely studied in ICU pa-tients and requires the use of a continu-ous infusion because of its very shortduration of action (35). The short dura-tion of action could be beneficial in se-lected patients requiring interruptionsfor neurologic examination (35).

Disease states, such as renal or hepaticinsufficiency may alter opioid and metab-olite elimination. Titration to the desiredresponse and assessment of the drug’sprolonged effect are necessary in all pa-tients. The elderly may have reduced opi-oid requirements (30, 31, 36–39).

Adverse effects of opioid analgesics arecommon and occur frequently in ICUpatients. Of greatest concern are respira-tory, hemodynamic, central nervoussystem, and gastrointestinal effects. Respi-ratory depression is a concern in spontane-ously breathing patients or those receivingpartial ventilatory support. Hypotensioncan occur in hemodynamically unstable pa-tients, hypovolemic patients, or those withelevated sympathetic tone (40). Opioid-mediated hypotension in euvolemic pa-tients is a result of the combination of sym-patholysis, vagally mediated bradycardia,and histamine release (when using codeine,morphine, or meperidine) (41, 42). Opioid-induced depression of the level of con-sciousness may cloud the clinical assess-ment of critically ill patients, andhallucinations may increase agitation insome patients. Gastric retention and ileusare common in critically ill patients, andintestinal hypomotility is enhanced by opi-oids (43, 44). Routine prophylactic use of astimulant laxative may minimize constipa-tion. Small-bowel intubation may beneeded for enteral nutrition because of gas-tric hypomotility (45). Opioids may in-crease intracranial pressure with traumaticbrain injury, although the data are incon-sistent and the clinical significance is un-known (46–48).

Opioid Administration Techniques.Preventing pain is more effective thantreating established pain. When patientsare administered drugs on an “as needed”basis, they may receive less than the pre-scribed dose and encounter significant de-lays in treatment, although the impact on

patient outcome has not been well docu-mented (49). Analgesics should be admin-istered on a continuous or scheduled inter-mittent basis, with supplemental bolusdoses as required (17). Intravenous admin-istration usually requires lower and morefrequent doses than intramuscular admin-istration to titrate to patient comfort. In-tramuscular administration is not recom-mended in hemodynamically unstablepatients because of altered perfusion andvariable absorption. When a continuous in-fusion was used, a protocol incorporatingdaily awakening from analgesia and seda-tion allowed more effective analgesic titra-tion and a lower total dose of morphine(50). Daily awakening was associated with ashorter duration of ventilation and ICU stay(50). A pain management plan and therapygoal should be established for each patientand reevaluated as the clinical conditionchanges. An algorithm that illustrates thepotential use of opioid analgesics for me-chanically ventilated patients is shown inFigure 1. Analgesic orders should be writ-ten to allow titration to achieve the analge-sic goal and to balance the potential impactof adverse effects.

In noncritically ill patients, patient-controlled analgesia (PCA) has been re-ported to result in stable drug concentra-tions, a good quality of analgesia, lesssedation, less opioid consumption, and po-tentially fewer adverse effects, including re-spiratory complications (10, 51). In addi-tion, a basal rate or continuous infusionmode can be used for consistent analgesiaduring sleep. Patient selection is importantwhen PCA is used, and particular attentionshould be paid to the patient’s cognition,hemodynamic reserve, and previous opioidexposure. PCA devices can also be used fornurse-controlled analgesia. The elimina-tion of paperwork can improve the timeli-ness of analgesic administration.

Fentanyl may also be administered viaa transdermal patch in hemodynamicallystable patients with more chronic analge-sic needs. The patch provides consistentdrug delivery, but the extent of absorp-tion varies depending on the permeabil-ity, temperature, perfusion, and thick-ness of the skin. There is a largeinterpatient variability in peak plasmaconcentrations. Fentanyl patches are nota recommended modality for acute anal-gesia because of their 12–24-hour delayto peak effect and similar lag time tocomplete offset once the patch is re-moved. Breakthrough pain should betreated with rapid-acting agents.

The use of a reversal agent, such as


naloxone, is not recommended after pro-longed analgesia, because it can inducewithdrawal and may cause nausea, cardiacstress, and arrhythmias. Analgesics withagonist-antagonist action, such as nalbu-phine, butorphanol, and buprenorphine,can also elicit withdrawal symptoms andshould be avoided during prolonged opioiduse.

Recommendations: A therapeutic planand goal of analgesia should be estab-lished for each patient and communi-cated to all caregivers to ensure con-sistent analgesic therapy. (Grade ofrecommendation � C)

If intravenous doses of an opioid anal-gesic are required, fentanyl, hydro-morphone, and morphine are the rec-ommended agents. (Grade ofrecommendation � C)

Scheduled opioid doses or a continuousinfusion is preferred over an “as needed”regimen to ensure consistent analgesia.A PCA device may be utilized to deliveropioids if the patient is able to under-stand and operate the device. (Grade ofrecommendation � B)

Fentanyl is preferred for a rapid onset ofanalgesia in acutely distressed patients.(Grade of recommendation � C)

Fentanyl or hydromorphone are pre-ferred for patients with hemodynamicinstability or renal insufficiency.(Grade of recommendation � C)

Morphine and hydromorphone arepreferred for intermittent therapy be-

cause of their longer duration of effect.(Grade of recommendation � C)

Nonopioid Analgesics. NSAIDs provideanalgesia via the nonselective, competi-tive inhibition of cyclooxygenase (COX), acritical enzyme in the inflammatory cas-cade. NSAIDs have the potential to causesignificant adverse effects, including gas-trointestinal bleeding, bleeding second-ary to platelet inhibition, and the devel-opment of renal insufficiency. Patientswith hypovolemia or hypoperfusion, theelderly, and those with preexisting renalimpairment may be more susceptible toNSAID-induced renal injury (52, 53). Pro-longed use (more than five days) of ketoro-lac has been associated with a two-fold in-crease in the risk of renal failure and anincreased risk of gastrointestinal and oper-ative-site bleeding (54, 55). NSAIDs shouldnot be administered to patients withasthma and aspirin sensitivity.

Administration of NSAIDs may reduceopioid requirements, although the anal-gesic benefit of NSAIDs has not been sys-tematically studied in critically ill pa-tients. Many oral agents are available, andibuprofen and naproxen are available inliquid form. Ketorolac is currently theonly parenteral NSAID. The safety of ke-torolac administration in patients withsevere renal insufficiency or those under-going dialysis has not been determined.

The role, if any, of the more selectiveCOX-2 inhibitors in the critically ill re-mains unknown. Selective COX-2 inhib-iting agents cause less gastrointestinal

irritation with long-term use than tradi-tional NSAIDs (56). The slow onset ofaction of some agents may decrease theirutility for acute pain management.

Acetaminophen is an analgesic used totreat mild to moderate pain. In combina-tion with an opioid, acetaminophen pro-duces a greater analgesic effect than higherdoses of the opioid alone (57). The role ofacetaminophen in critical care is limited torelieving mild pain or discomfort, such asthat associated with prolonged bed rest oruse as an antipyretic. Care must be taken toavoid excessive and potentially hepatotoxicdoses, especially in patients with depletedglutathione stores resulting from hepaticdysfunction or malnutrition. Acetamino-phen should be maintained at less than 2 gper day for patients with a significant his-tory of alcohol intake or poor nutritionalstatus and less than 4 g per day for others(Table 1) (58).

Recommendations: NSAIDs or acet-aminophen may be used as adjuncts toopioids in selected patients. (Grade ofrecommendation � B)

Ketorolac therapy should be limited to amaximum of five days, with close mon-itoring for the development of renal in-sufficiency or gastrointestinal bleeding.Other NSAIDs may be used via the en-teral route in appropriate patients.(Grade of recommendation � B)

SEDATION

The indications for sedative agents arenot well defined. Sedatives are common

Table 1. Pharmacology of selected analgesics (1, 17, 30–32)

AgentEquianalgesic

Dose (i.v.) Half-life Metabolic Pathway Active Metabolites (Effect) Adverse Effects

Fentanyl 200 �g 1.5–6 hr Oxidation No metabolite, parent accumulates Rigidity with high dosesHydromorphone 1.5 mg 2–3 hr Glucuronidation None . . .Morphine 10 mg 3–7 hr Glucuronidation Yes (sedation, especially in renal

insufficiency)Histamine release

Meperidine 75–100 mg 3–4 hr Demethylation andhydroxylation

Yes (neuroexcitation, especially inrenal insufficiency or high doses)

Avoid with MAOIsc and SSRIsd

Codeine 120 mg 3 hr Demethylation andglucuronidation

Yes (analgesia, sedation) Lacks potency, histamine release

Remifentanil . . . 3–10 min Plasma esterase None . . .Ketorolac . . . 2.4–8.6 hr Renal None Risk of bleeding, GI and renal

adverse effects

Ibuprofen . . . 1.8–2.5 hr Oxidation None Risk of bleeding, GI and renaladverse effects

Acetaminophen . . . 2 hr Conjugation . . . . . .

aMore frequent doses may be needed for acute pain management in mechanically ventilated patients.bCost based on 2001 average wholesale price.cMAOIs � monoamine oxidase inhibitors.dSSRIs � selective serotonin-reuptake inhibitors.


adjuncts for the treatment of anxiety andagitation. The causes of anxiety in criticallyill patients are multifactorial and are likelysecondary to an inability to communicateamid continuous noise (alarms, personnel,and equipment), continuous ambient light-ing, and excessive stimulation (inadequateanalgesia, frequent vital signs, reposition-ing, lack of mobility, and room tempera-ture). Sleep deprivation and the circum-stances that led to an ICU admission mayincrease patient anxiety, affecting up to50% of ICU patients (38, 59). Efforts toreduce anxiety, including frequent reorien-tation, maintenance of patient comfort,provision of adequate analgesia, and opti-mization of the environment, may be sup-plemented with sedatives. Some patientswith respiratory failure require sedation tofacilitate mechanical ventilation, althoughsedation should not be used in lieu of ap-propriate ventilation strategies.

Agitation is common in ICU patientsof all ages, occurring at least once in 71%of patients in a medical-surgical ICU (38).Agitation can be caused by multiple fac-tors, such as extreme anxiety, delirium,adverse drug effects, and pain (38). How-ever, not all patients with anxiety willexhibit agitation; some patients may befearful, anxious, and withdrawn. Whenpatients exhibit signs of anxiety or agita-tion, the first priority is to identify andtreat any underlying physiological distur-bances, such as hypoxemia, hypoglyce-mia, hypotension, pain, and withdrawalfrom alcohol and other drugs. The prev-alence of drug and alcohol abuse in thegeneral population is high, and thesesubstances are commonly associated withtraumatic injury (60–62). Patients in theICU should be assessed for symptoms of

intoxication or withdrawal upon admis-sion and for several weeks thereafter (60–62). When possible, patients should bequestioned about the use of herbal med-icines because these products may con-tribute to significant drug interactionsand adverse effects (63).

Recent studies have confirmed thatagitation may have a deleterious effect onpatients by contributing to ventilatordysynchrony, an increase in oxygen con-sumption, and inadvertent removal of de-vices and catheters (38, 64–67). Seda-tives reduce the stress response andimprove the tolerance of routine ICU pro-cedures (68). The use of sedatives tomaintain patient safety and comfort isoften essential to the ICU therapeuticcare plan. The sedation of mechanicallyventilated patients is often medically nec-essary and should be based on an individ-ualized assessment and the patient’sneeds. Sedatives should be administeredintermittently or on an “as needed” basisto determine the dose that will achievethe sedation goal. Sedatives, as outlinedin this guideline, are not intended to beused as a method of restraint and are notto be “used as a means of coercion, dis-cipline, convenience, or retaliation bystaff” (Federal regulation 42 CFR 482.13).It is important to consider this principlein order to follow the intent of the Cen-ters for Medicare and Medicaid Servicesregulation regarding restraints.

An analgesic may be the appropriateinitial therapy when pain is the suspectedcause of acute agitation. Although opi-oids may produce sedating effects, theydo not diminish awareness or provideamnesia for stressful events. Sedative-amnestic therapy is required to reliably

attain amnesia (69–72). Without amne-sia, many patients who recall their ICUstay report unpleasant or frighteningmemories, which may contribute to post-traumatic stress disorder (PTSD) symp-toms (21, 73). However, some patientshave vivid hypnogagic hallucinations(dreams just before loss of consciousness)with sedative-amnestic therapy (74). Assedation blunts explicit memory, thesehallucinations may be patients’ onlymemory of the ICU experience (75). Re-calling delusions, without memory of realevents, may also contribute to acutePTSD-related symptoms (75). Other datasuggest that PTSD may be experienced by4–15% of ICU survivors (76, 77). Amnes-tic sedatives may paradoxically contributeto agitation and disorientation becausepatients may not remember where theyare or why they are in the ICU.

Recommendation: Sedation of agi-tated critically ill patients should bestarted only after providing adequateanalgesia and treating reversible phys-iological causes. (Grade of recommen-dation � C)

Sedation Assessment

Subjective Assessment of Sedationand Agitation. Frequent assessment ofthe degree of sedation or agitation mayfacilitate the titration of sedatives to pre-determined endpoints (78–80). An idealsedation scale should provide data thatare simple to compute and record, accu-rately describe the degree of sedation oragitation within well-defined categories,guide the titration of therapy, and havevalidity and reliability in ICU patients.Many scales are available, but a true gold-standard scale has not been established(79). Several scales have construct valid-ity with good correlation between thescales’ measures and other measures ofsedation. None of the scales have beentested for their ability to detect a patient’sresponse to changes in sedative therapy,dosage, or withdrawal. However, a de-fined sedation goal, using the Ramsayscale and a protocol-driven sedation plan,was shown to reduce the duration of me-chanical ventilation and length of stay(80). The authors did not report otherpatient outcome measures relative to theadequacy of analgesia or sedation.

The Riker Sedation-Agitation Scale(SAS) was the first scale proven to bereliable and valid in critically ill adults(81, 82). SAS scores a patient’s level of

Table 1. (Continued)

Intermittent DoseaInfusion Dose Range

(Usual)Infusion Cost per day

70 kgb

0.35–1.5 �g/kg i.v. q 0.5–1 hr 0.7–10 �g/kg/hr 100 �g/h: $26.0010–30 �g/kg i.v. q 1–2 hr 7–15 �g/kg/hr 0.75 mg/hr: $5.00–11.000.01–0.15 mg/kg i.v. q 1–2 hr 0.07–0.5 mg/kg/hr 5 mg/hr: $3.50–12.00

Not recommended Not recommended . . .

Not recommended Not recommended . . .

. . . 0.6–15 �g/kg/hr 10 �g/kg/hr: $170.0015–30 mg i.v. q 6h, decrease if age

� 65 yr or wt � 50 kg or renalimpairment, avoid � 5 days use.

. . .

400 mg p.o. q 4–6 hr . . . . . .

325–650 mg p.o. q 4–6 hr, avoid� 4 g/day

. . . . . .


consciousness and agitation from a sev-en-item list describing patient behavior(Table 2). Excellent inter rater reliabilityhas been demonstrated and validity hasbeen shown with two other scales. TheMotor Activity Assessment Scale (MAAS),adapted from the SAS, has also been val-idated and shown reliable for use in crit-ically ill patients (83). The MAAS hasseven categories to describe patient be-haviors in response to stimulation (Table2). The Ramsay scale measures three lev-els of awake states and three levels ofasleep states (Table 2) (84). It has beenshown to have an acceptable interraterreliability compared with the SAS, buthas been criticized for its lack of cleardiscrimination and specific descriptors todifferentiate between the various levels(82, 85). Nevertheless, the Ramsay scalehas been used in many comparative seda-tion trials and is widely used clinically.The Vancouver Interaction and CalmnessScale (VICS) has also been validated for

the assessment of sedation in adult criti-cally ill patients (86). With the VICS scor-ing system, patients are assessed inde-pendently for the ability to interact andcommunicate and for their level of activ-ity or restlessness. The VICS has not beentested to identify optimal sedation end-points. Another scale, the Observer’s As-sessment of Alertness/Sedation Scale, isoften used in the operating room butlacks the ability to assess agitation andhas never been tested in the ICU (87). TheCOMFORT scale has been extensivelytested and applied in the ICU environ-ment, but only in children (88).

The appropriate target level of seda-tion will primarily depend on a patient’sacute disease process and any therapeuticand supportive interventions required. Acommon target level of sedation in theICU is a calm patient that can be easilyaroused with maintenance of the normalsleep-wake cycle, but some may requiredeep levels of sedation to facilitate me-

chanical ventilation. The desired level ofsedation should be defined at the start oftherapy and reevaluated on a regular ba-sis as the clinical condition of the patientchanges. Regimens should be writtenwith the appropriate flexibility to allowtitration to the desired endpoint, antici-pating fluctuations in sedation require-ments throughout the day.

Objective Assessment of Sedation. Ob-jective testing of a patient’s level of seda-tion may be helpful during very deep se-dation or when therapeutic neuro-muscular blockade masks observable be-havior. Vital signs, such as blood pressureand heart rate, are not specific or sensi-tive markers of the level of sedationamong critically ill patients. Tools uti-lized in objective assessment includeheart rate variability and lower-esopha-geal contractility, but most are based on apatient’s electroencephalogram (EEG).The raw EEG signal has been manipu-lated by using several devices to simplify

Figure 1. Algorithm for the sedation and analgesia of mechanically ventilated patients. This algorithm is a general guideline for the use of analgesics andsedatives. Refer to the text for clinical and pharmacologic issues that dictate optimal drug selection, recommended assessment scales, and precautions forpatient monitoring. Doses are approximate for a 70-kg adult. IVP � intravenous push.


bedside interpretation and improve reliabil-ity. For example, the bispectral index (BIS)uses a digital scale from 100 (completelyawake) to 0 (isoelectric EEG) (89). Most ofthe literature about the use of BIS in theoperating room supports strong agreementbetween BIS and patient recall or level ofhypnosis (90). Elective surgery patients re-ceiving sedatives have shown a strong in-verse correlation between hypnotic drugeffect and BIS (91, 92).

Although the BIS may be a promisingtool for the objective assessment of seda-tion or hypnotic drug effect, it has limita-tions in the ICU environment (93–95). BISscores may vary between patients at thesame subjective level of sedation, and sub-jective scales may be more reproducibleduring light sedation (93, 94). Muscle-based electrical activity may artificially ele-vate BIS scores if the patient has not re-ceived neuromuscular blockade (94). A newversion of BIS software is being tested forimproved applicability in measuring ICU

sedation. BIS has not been tested in pa-tients with metabolic impairments orstructural abnormalities of the brain. Stud-ies have not compared the patient out-comes of using BIS versus subjective scales.Although BIS is likely to be useful whenpatients are deeply comatose or under neu-romuscular blockade, routine use of thisdevice cannot be recommended until thevalue and validity are confirmed.

Recommendations: A sedation goal orendpoint should be established and reg-ularly redefined for each patient. Regu-lar assessment and response to therapyshould be systematically documented.(Grade of recommendation � C)

The use of a validated sedation assess-ment scale (SAS, MAAS, or VICS) isrecommended. (Grade of recommen-dation � B)

Objective measures of sedation, such asBIS, have not been completely evaluated

and are not yet proven useful in the ICU.(Grade of recommendation � C)

Sedation Therapy

Benzodiazepines. Benzodiazepines aresedatives and hypnotics that block theacquisition and encoding of new informa-tion and potentially unpleasant experi-ences (anterograde amnesia) but do notinduce retrograde amnesia. Althoughthey lack any analgesic properties, theyhave an opioid-sparing effect by moder-ating the anticipatory pain response (96,97). Benzodiazepines vary in their po-tency, onset and duration of action, up-take, distribution, metabolism, and pres-ence or absence of active metabolites(Table 3). Patient-specific factors, such asage, concurrent pathology, prior alcoholabuse, and concurrent drug therapy, af-fect the intensity and duration of activityof benzodiazepines, requiring individual-ized titration. Elderly patients exhibit

Table 2. Scales used to measure sedation and agitation

Score Description Definition

Riker Sedation-Agitation Scale (SAS) (82)7 Dangerous agitation Pulling at endotracheal tube (ETT), trying to remove catheters, climbing over

bedrail, striking at staff, thrashing side-to-side6 Very agitated Does not calm despite frequent verbal reminding of limits, requires physical

restraints, biting ETT5 Agitated Anxious or mildly agitated, attempting to sit up, calms down to verbal

instructions4 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

commandsMotor Activity Assessment Scale (MAAS) (83)

6 Dangerously agitated No external stimulus is required to elicit movement and patient is uncooperativepulling at tubes or catheters or thrashing side to side or striking at staff ortrying to climb out of bed and does not calm down when asked

5 Agitated No external stimulus is required to elicit movement and attempting to sit up ormoves limbs out of bed and does not consistently follow commands (e.g., willlie down when asked but soon reverts back to attempts to sit up or move limbsout of bed)

4 Restless and cooperative No external stimulus is required to elicit movement and patient is picking atsheets or tubes or uncovering self and follows commands

3 Calm and cooperative No external stimulus is required to elicit movement and patient is adjustingsheets or clothes purposefully and follows commands

2 Responsive to touch or name Opens eyes or raises eyebrows or turns head toward stimulus or moves limbswhen touched or name is loudly spoken

1 Responsive only to noxious stimulusa Opens eyes or raises eyebrows or turns head toward stimulus or moves limbs withnoxious stimulus

0 Unresponsive Does not move with noxious stimulusRamsay Scale (84)

1 Awake Patient anxious and agitated or restless or both2 Patient cooperative, oriented and tranquil3 Patient responds to commands only4 Asleep A brisk response to a light glabellar tap or loud auditory stimulus5 A sluggish response to a light glabellar tap or loud auditory stimulus6 No response to a light glabellar tap or loud auditory stimulus

aNoxious stimulus � suctioning or 5 seconds of vigorous orbital, sternal, or nail bed pressure.


slower clearance of benzodiazepines ortheir active metabolites and have a largervolume of drug distribution, contributingto a marked prolongation of elimination(111). Compromised hepatic or renal func-tion may slow the clearance of benzodiaz-epines or their active metabolites. Induc-tion or inhibition of hepatic or intestinalenzyme activity can alter the oxidative me-tabolism of most benzodiazepines (112).

Benzodiazepine therapy should be ti-trated to a predefined endpoint, often re-quiring a series of loading doses. Hemo-dynamically unstable patients mayexperience hypotension with the initia-tion of sedation. Maintenance of sedationwith intermittent or “as needed” doses ofdiazepam, lorazepam, or midazolam maybe adequate to accomplish the goal ofsedation (78). However, patients requir-ing frequent doses to maintain the de-sired effect may benefit from a continu-ous infusion by using the lowest effectiveinfusion dose. Continuous infusionsmust be used cautiously, as accumulationof the parent drug or its active metabo-lites may produce inadvertent overseda-tion. Frequent reassessment of a patient’ssedation requirements and active taper-ing of the infusion rate can prevent pro-longed sedative effects (80). However,awakening times after several days of se-dation may be quite unpredictable inclinical use. In contrast, tolerance to ben-zodiazepines may occur within hours toseveral days of therapy, and escalatingdoses of midazolam have been reported(113, 114). While not well described inthe literature, paradoxical agitation hasalso been observed during light sedationand may be the result of drug-inducedamnesia or disorientation.

Diazepam has been shown to provide

rapid onset and awakening after singledoses (Table 3) (78, 115). Because of itslong-acting metabolites, a prolonged du-ration of sedative effect may occur withrepeated doses, but this may be accept-able for long-term sedation (78). Loraz-epam has a slower onset but fewer poten-tial drug interactions because of itsmetabolism via glucuronidation (Table 3)(98, 112). The slow onset makes loraz-epam less useful for the treatment ofacute agitation. Maintenance of sedationcan be accomplished with intermittent orcontinuous intravenous administration.Lorazepam has an elimination half-life of12–15 hours, so an infusion is not readilytitratable. Loading doses given by i.v.push should be used initially with rela-tively fixed infusion rates. Lorazepam in-fusions should be prepared using the 2mg/mL injection and diluted to a concen-tration of 1 mg/mL or less and mixed in aglass bottle (116, 117). Despite these pre-cautions, precipitation may develop(118). An alternative is to administer un-diluted lorazepam as an infusion using aPCA device (78). The lorazepam solventspolyethylene glycol (PEG) and propyleneglycol (PG) have been implicated as thecause of reversible acute tubular necro-sis, lactic acidosis, and hyperosmolarstates after prolonged high-dose infu-sions. The dosing threshold for this effecthas not been prospectively defined, butthese case reports described doses thatexceeded 18 mg/hr and continued forlonger than four weeks and higher doses(�25 mg/hr) continuing for hours to days(119–121). It seems prudent to avoid dosesof this magnitude. Alternatively, lorazepamand diazepam may be administered via theenteral route in tablet or liquid form (122).Large doses of liquid lorazepam (i.e., 60 mg

of 2 mg/mL every six hours) may lead todiarrhea because of the high PEG and PGcontent (123).

Midazolam has a rapid onset and shortduration with single doses, similar to di-azepam (Table 3) (115). The rapid onsetof midazolam makes it preferable fortreating acutely agitated patients. Accu-mulation and prolonged sedative effectshave been reported in critically ill pa-tients using midazolam who are obese orhave a low albumin level or renal failure(99–103). Prolonged sedative effects mayalso be caused by the accumulation of anactive metabolite, alpha-hydroxymidazo-lam, or its conjugated salt, especially inpatients with renal insufficiency (101–105). Significant inhibition of midazolammetabolism has been reported withpropofol, diltiazem, macrolide antibiot-ics, and other inhibitors of cytochromeP450 isoenzyme 3A4, which could influ-ence the duration of effect (107, 108,112). Daily discontinuation of midazolaminfusions (wake up) with retitration to aRamsay scale endpoint reduced midazo-lam requirements and was associatedwith a reduction in the duration of me-chanical ventilation and length of ICUstay (50). However, the patients in thistrial were off of midazolam for an averageof 5.3 hours per day, so this researchtechnique may be difficult to implement.Patients should be closely monitored forself-extubation or the removal of othermonitoring devices during the dailyawakening sessions.

The routine use of a benzodiazepineantagonist, such as flumazenil, is not rec-ommended after prolonged benzodiaz-epine therapy because of the risks ofinducing withdrawal symptoms and in-creasing myocardial oxygen consumption

Table 3. Pharmacology of selected sedatives (1, 30–32, 98–110)

AgentOnset After

i.v. DoseHalf-life of Parent

Compound Metabolic Pathway Active Metabolite Unique Adverse Effects

Diazepam 2–5 min 20–120 hr Desmethylation andhydroxylation

Yes (prolonged sedation) Phlebitis

Lorazepam 5–20 min 8–15 hr Glucuronidation None Solvent-related acidosis/renalfailure in high doses

Midazolam 2–5 min 3–11 hr Oxidation Yes (prolonged sedation,especially with renalfailure)

Propofol 1–2 min 26–32 hr Oxidation None Elevated triglycerides, painon injection

Haloperidol 3–20 min 18–54 hr Oxidation Yes (EPS)c QT interval prolongation

aMore frequent doses may be needed for management of acute agitation in mechanically ventilated patients.bCost based on 2001 average wholesale price.cEPS � extrapyramidal symptoms.


with as little as 0.5 mg of flumazenil(124). An i.v. dose of flumazenil 0.15 mgis associated with few withdrawal symp-toms when administered to patients re-ceiving midazolam infusions (125). Ifflumazenil is used to test for prolongedsedation after several days of benzodiaz-epine therapy, a single low dose is recom-mended.

Propofol. Propofol is an intravenous,general anesthetic agent. However, sedativeand hypnotic properties can be demon-strated at lower doses. Compared with ben-zodiazepines, propofol produces a similardegree of amnesia at equisedative doses involunteers (69). In a clinical trial of ICUpatients, propofol did not produce amnesiaas often as midazolam (70). Like the ben-zodiazepines, propofol has no analgesicproperties.

Propofol has a rapid onset and shortduration of sedation once discontinued(Table 3). While most of the early litera-ture documents the comparatively rapidresolution of sedation after propofol infu-sions, a slightly longer recovery has beenreported after more than 12 hours of in-fusion (110, 126). No changes in kineticparameters have been reported in pa-tients with renal or hepatic dysfunction.

Propofol is available as an emulsion ina phospholipid vehicle, which provides1.1 kcal/mL from fat and should becounted as a caloric source. Long-term orhigh-dose infusions may result in hyper-triglyceridemia (127–129). Other adverseeffects most commonly seen with propo-fol include hypotension, bradycardia, andpain upon peripheral venous injection.The hypotension is dose related and morefrequent after bolus dose administration.Elevation of pancreatic enzymes has beenreported during prolonged infusions ofpropofol (130, 131). Pancreatitis has beenreported following anesthesia with propo-fol, although a causal relationship has

not been established (132). Prolonged use(�48 hours) of high doses of propofol(�66 �g/kg/min infusion) has been asso-ciated with lactic acidosis, bradycardia,and lipidemia in pediatric patients anddoses �83 �g/kg/min have been associ-ated with an increased risk of cardiacarrest in adults (133, 134). The adults athighest risk for cardiac complications re-ceived �100 �g/kg/min infusion of a 2%propofol solution to achieve deep sedationafter neurologic injury (134). FDA has spe-cifically recommended against the use ofpropofol for the prolonged sedation of pe-diatric patients. Patients receiving propofolshould be monitored for unexplained met-abolic acidosis or arrhythmias.

Alternative sedative agents should beconsidered for patients with escalating va-sopressor or inotrope requirements or car-diac failure during high-dose propofol infu-sions.

Propofol requires a dedicated i.v. cath-eter when administered as a continuousinfusion because of the potential for drugincompatibility and infection. Improperaseptic technique with propofol in theoperating room has led to nosocomialpostoperative infection (135). However, aclinically relevant incidence of infectiouscomplications has not been reported withICU use (136). The manufacturers sug-gest that propofol infusion bottles andtubing should hang no more than 12hours and solutions transferred from theoriginal container should be discardedevery 6 hours. A preservative has beenadded to propofol to decrease the poten-tial for bacterial overgrowth in case thevial would become contaminated. One ofthe propofol formulations contains edeticacid (Diprivan, AstraZeneca) and themanufacturer recommends a drug holi-day after more than seven days of infu-sion to minimize the risk of trace ele-ment abnormalities. Another product

(propofol, Gensia Sicor) contains sodiummetabisulfite, which may produce aller-gic reactions in susceptible patients. Sul-fite sensitivity occurs more frequently inpatients with asthma.

While propofol appears to possess an-ticonvulsant activity, excitatory phenom-ena, such as myoclonus, have been ob-served. There are several case reports andsmall, uncontrolled studies describingthe efficacy of propofol in refractory sta-tus epilepticus (after traditional treat-ment regimens have failed or are nottolerated) and electroconvulsive shocktherapy (137, 138). Case reports have alsodescribed roles for propofol in deliriumtremens refractory to high-dose benzodi-azepine therapy (139).

Propofol has been used to sedate neu-rosurgical patients to reduce elevated in-tracranial pressure (ICP) (140, 141). Therapid awakening from propofol allows in-terruption of the infusion for neurologicassessment. Propofol may also decreasecerebral blood flow and metabolism.Propofol and morphine produced im-proved control of ICP compared withmorphine alone in the treatment of se-vere traumatic brain injury (TBI) (141).Propofol reduced elevated ICP more ef-fectively than fentanyl following severeTBI (142). High doses of propofol shouldbe used cautiously in this setting (134).Propofol infusions used to reduce ele-vated ICP may need to be continuedlonger than usually recommended forroutine sedation (132).

Central �-Agonists. Clonidine has beenused to augment the effects of general an-esthetics and narcotics and to treat drugwithdrawal syndromes in the ICU (143,144). The more selective �-2 agonist,dexmedetomidine, was recently approvedfor use as a sedative with analgesic-sparingactivity for short-term use (�24 hours) inpatients who are initially receiving me-chanical ventilation. Patients remain se-dated when undisturbed, but arouse readilywith gentle stimulation. Dexmedetomidinereduces concurrent analgesic and sedativerequirements and produces anxiolytic ef-fects comparable to benzodiazepines (145–148). Rapid administration of dexmedeto-midine may produce transient elevations inblood pressure. Patients maintained ondexmedetomidine may develop bradycardiaand hypotension, especially in the presenceof intravascular volume depletion or highsympathetic tone. The role of this newagent in the sedation of ICU patients re-mains to be determined.


Intermittenti.v. Dosea

Infusion Dose Range(Usual) Cost per day 70 kg patientb

0.03–0.1 mg/kgq 0.5–6 hr

. . . 20 mg q 4 hr: $5.00–20.50

0.02–0.06 mg/kgq 2–6 hr

0.01–0.1 mg/kg/hr 48 mg/day: $55.00

0.02–0.08 mg/kgq 0.5–2 hr

0.04–0.2 mg/kg/hr 6 mg/hr: $65.00–309.00

. . . 5–80 �g/kg/min 50 �g/kg/min: $235.00–375.00

0.03–0.15 mg/kgq 0.5–6 hr

0.04–0.15 mg/kg/hr 10 mg q 6 h: $62.00–65.00


Sedative Selection

Acute agitation arises from a variety ofetiologies, including pain. A short-actingopioid analgesic, such as fentanyl, mayprovide immediate sedation and patientcomfort; however, fentanyl has not beencompared with other sedatives in a con-trolled trial. Midazolam and diazepamalso have a rapid onset of sedation (115).Propofol has a rapid onset, but hypoten-sion and infusion-site pain can resultfrom bolus dose administration. Cautioususe of sedatives is recommended for pa-tients not yet intubated because of therisk of respiratory depression.

Comparative trials of prolonged seda-tion have been performed in a variety ofcritical care settings. Many were sup-ported by pharmaceutical industry re-search grants; as a result, newer productshave been evaluated more frequently.

Outcome is usually described in terms ofthe speed of onset, ability to maintain thetarget level of sedation, adverse effects,time required for awakening, and abilityto wean from mechanical ventilation.Most of the prospective, randomized tri-als are experimentally flawed becausethey are unblinded, use uncontrolledamounts of opioids, and exclude patientswith obesity or renal or hepatic insuffi-ciency. This limits their general applica-bility. There is a need for more large,high-quality, randomized trials of the ef-fectiveness of different sedative agents(149). Most of the trials used a Ramsayscale for assessment, so the depth of se-dation can generally be compared amongthe trials. The trials are summarized inTables 4–6.

Duration of Therapy. Short-term(�24 hours), randomized, open-label tri-

als of sedation have compared propofoland midazolam most often (eight of ninetrials) (Table 4). An opioid was availableto all patients. Awakening times for pa-tients taking propofol ranged from 1 to105 minutes versus 1 to 405 minutes forpatients receiving midazolam (128, 150–157). Time to extubation has also beencompared, but other variables may influ-ence this outcome measure. Clinically,these agents produced similar outcomesfollowing �24 hours of infusion (Table 4).

An intermediate duration of sedation(one to three days) was reported in threerandomized open-label trials (Table 5).Propofol and midazolam were comparedfor sedation of medical ICU patients and amixed medical-surgical ICU populationwith respiratory failure (157, 158). Pa-tients receiving propofol had statisticallymore predictable awakening times than

Table 4. Clinical trials with less than 24 hours of sedationa

Level ofEvidence

(Reference)Population Type

(No. Patients) Exclusion Criteria Trial Design Drugs Mean Dosage

1 (150) MICU/SICU (100) Obese, head injury,NMBA use

Multicenter,open-label

PropofolMidazolamMorphine: to all patients

Propofol: 1.77 mg/kg/hrMidazolam: 0.1 mg/kg/hr

2 (128) MICU/SICU (88total, 40short-termsedation)

Neurologic injury,ongoing NMBA use

Open-label PropofolMidazolamMorphine: to all patients


2 (151) CABG (30) Obese, renal or hepaticinsufficiency

Open-label PropofolMidazolamSufentanil: to all patients

Propofol: 2.71 � 1.13 mg/kg/hrMidazolam: 0.09 � 0.03 mg/kg/hr

1 (152) CABG (84) Renal or hepaticinsufficiency

Open-label PropofolMidazolamMorphine: p.r.n.


1 (153) SICU (60) None Open-label,consecutivepatients

PropofolMidazolamMorphine: p.r.n.

Propofol: 114.8 mg/hr (80.1 � 21.1kg)

Midazolam: 2.1 � 1.3 mg/hr(74.2 � 24 kg)

1 (154) MICU/SICU (61%trauma)

Cardiac insufficiency,neurosurgical orunstable

Multicenter,open-label

LorazepamMidazolamMorphine: p.r.n.

Lorazepam: 1.6 � 0.1 mg i.v. push,Midazolam: 14.4 � 1.2 mg infusion

over 8 hrs1 (155) Cardiac surgery

(41)Renal, hepatic, or cardiac

failureDouble-blind Propofol

MidazolamMorphine: p.r.n.

Propofol: 0.64 � .17 mg/kg/hrMidazolam: 0.015 � 0.001 mg/kg/hr

1 (156) CABG (75) Renal failure, neurologichistory, hepatic failure,cardiovasculardysfunction

Double-blind PropofolMidazolamBoth Propofol and

MidazolamMorphine: to all patients

Propofol: 1.2 � 0.03 mg/kg/hrMidazolam: 0.08 � 0.01 mg/kg/hrPropofol and Midazolam:

Propofol: 0.22 � 0.03 mg/kg/hr,Midazolam: 0.02 mg/kg/hr

2 (157) MICU/SICU (99) Neurosurgery, coma,seizures

Multicenter,open-label,intention-to-treatanalysis

PropofolMidazolamOpioids: p.r.n.

Actual doses not specified

aMICU � medical intensive care unit, SICU � surgical intensive care unit, NMBA � neuromuscular blocking agent, GCS � Glasgow Coma Score,CABG � coronary artery bypass graft.


patients receiving midazolam in both tri-als. Clinically, this time difference wasnot as significant and did not producemore rapid discharge from the ICU (157).In a three-way comparison of midazolam,lorazepam, and propofol infusions for se-dation of surgical ICU patients, the au-thors concluded that lorazepam was thepreferred agent in this population (118).Overall, these agents were similar in thelevels of sedation provided, the time re-quired to achieve adequate sedation, andthe number of dose adjustments per day.However, midazolam produced adequatesedation during a greater percentage oftime while propofol was associated withmore undersedation and lorazepam withmore oversedation. Morphine was pro-vided on an as needed basis and the av-erage dose was similar in all threegroups. Awakening times were not re-ported. Precipitation of lorazepam infu-sions was reported (118).

Nine open-label, randomized trials com-paring long-term sedation (more thanthree days) were reviewed in these guide-lines (Table 6) (70, 127–129, 157, 159–163). Most of the trials compared propofolwith midazolam. All trials included opioidtherapy, although administration was notcontrolled. Most studies used a Ramsayscale for patient assessment. In these trials,propofol consistently produced more rapidawakening than midazolam with a statisti-cal and, probably, a clinical difference (70,127–129, 160, 161). Propofol patients awak-ened and were extubated in 0.25–4 hourswhile midazolam patients required 2.8–10hours to awaken and up to 49 hours forextubation (Table 6) (157). The greatest dif-ference in time to awakening was seenwhen a deep level of sedation was the goalof therapy (Ramsay level 4–5). Patients re-ceiving propofol awakened from deep seda-tion significantly faster than those receiv-ing midazolam (127–129). Lorazepam and

midazolam were also compared for long-term sedation (159, 162). One of thesestudies used a double-blind study design(162). There was no statistically significantdifference in awakening time between theseagents when titrated to similar levels ofsedation, although the awakening times as-sociated with lorazepam appeared to bemore predictable.

Sedative Comparison. Four trialscompared lorazepam with midazolam(118, 154, 159, 162). Intermittent loraz-epam doses produced sedation compara-ble to a midazolam infusion during aneight-hour observation period (154).Both lorazepam and midazolam have thepotential to cause accumulation and pro-longed drug effects or oversedation if ad-ministered excessively via continuous in-fusion, especially when deep levels ofsedation are attempted (118, 159). How-ever, a rigorous protocol of assessmentand titration of lorazepam infusions to


Actual Duration Sedation Endpoint or Goal Outcome Measure Significance or Conclusion

Propofol: 20.2 hrMidazolam: 21.3 hr

Ramsay level 2–4 Awakening time: most awake at end ofinfusion, longest: Propofol: 105 min,Midazolam: 405 min

No statistical analysis reported


Ramsay level 2–5 and modifiedGCS

Time to extubation: Propofol: 0.3 hrMidazolam: 2.5 � 0.9 hr

p � 0.05, Propofol more rapidextubation


Ramsay level 5 Time to extubation:Propofol: 250 � 135 minMidazolam: 391 � 128 min

p � 0.01, Propofol more rapidextubation


Ramsay level 3 Time to extubation:Propofol: 4.3 hrMidazolam: 3.5 hr

Not statistically significant

16 hr observation, totalsedation duration notdefined

Ramsay level 3 and response tostimulation

Postsedation score at 5, 30, 60, and 90min: Propofol scores lower at 5 and30 min

p � 0.05, Midazolam moreheavily sedated

8 hr Multiple scales: anxiety,amnesia, pain, GCS, Ramsaylevel 3

Adequacy of sedation Comparable sedation scores

Propofol: 4 hrMidazolam: 4 hr

Ramsay level 2–4 Awakening time:Propofol: 88.6 � 51 minMidazolam: 93.8 � 59.4 min



Propofol andMidazolam: 14.7 hr

Modified GCS Time to extubation:Propofol: 0.9 � 0.3 hrMidazolam: 2.3 � 0.8 hrPropofol and Midazolam: 1.2 � 0.6 hr

p � 0.01, Midazolam vs.Propofol or Propofol andMidazolam

�24 hr—post hocstratum

Propofol: n � 21Midazolam: n � 26

Ramsay scale, level wasspecified daily

Time to extubation:Propofol: 5.6 hrMidazolam: 11.9 hr

p � 0.029 Overall: Propofol:60.2% of time at targetRamsay score Midazolam:44% of time at targetRamsay score, p � 0.05

aMICU � medical intensive care unit, SICU � surgical intensive care unit, NMBA � neuromuscular blocking agent, GCS � Glasgow Coma Score,CABG � coronary artery bypass graft.


moderate levels of sedation produced aless variable awakening time with loraz-epam than with midazolam, although theabsolute difference in awakening timewas not statistically significant (159). Anurse-managed sedation protocol, whichincluded the active titration of lorazepaminfusions to a defined endpoint, avoided aprolonged sedative effect compared withphysician management of infusion rates(80). A blinded trial of lorazepam versusmidazolam found that the lorazepam in-fusions were easier to manage than mi-dazolam infusions because fewer dose ad-justments were required to maintain thedesired level of sedation (162). In thistrial, wide inter- and intrapatient vari-ability was noted between sedative plasmaconcentrations and the Ramsay score. Nodifference was noted in patient recoverywhen patients were evaluated for 24hours after the end of the infusion. Awak-ening times were not reported in two ofthe other trials (118, 154).

When titrated to a standard endpoint,midazolam and propofol provide compara-ble levels of sedation with a similar onset ofeffect (128, 150–153, 155–157). As shownin Table 4, there is generally no statisticalor clinical difference in awakening timesbetween propofol and midazolam whenused for short-term sedation. Data fromlonger trials of sedation (more than 72hours) suggest that propofol is associatedwith more reliable and rapid awakening,both statistically and clinically, than mida-zolam (106, 127, 128, 156, 158, 160). Fol-lowing long-term sedation, propofol awak-ening times ranged from 0.25 to 2.5 hoursand midazolam awakening times ranged

from 2.8 to 30 hours (Table 6). One center’scomparison of two concentrations ofpropofol (1% and 2%) versus midazolamused in trauma patients has shown thatmidazolam provides deep levels of sedationmore reliably than propofol, but the awak-ening times were much longer with mida-zolam (127, 163). More patients receiving1% propofol experienced failure because ofelevated triglycerides, but the 2% propofolgroup experienced failure because of inad-equate sedation. Failure of propofol to pro-vide adequate sedation of trauma patientswas reported elsewhere, although an expla-nation was not apparent (118).

Economic Comparison. Several ofthese studies presented limited phar-macoeconomic data. In most reports, thesedative costs were the only costs consid-ered (cost minimization) (118, 154–156,162). Some studies included a portion ofthe costs for ICU patient care (128, 161).A sedative with a low acquisition cost maybe cited as the least expensive agent forprolonged sedation (e.g., lorazepam) (1,32, 118). A complete economic analysisshould consider costs associated with theevaluation and treatment of sedation-induced adverse effects (e.g., prolongedsedation, infection, and hypertriglyceri-demia), therapy failures (additionalagents or high doses required), and drugpreparation and administration costs(precipitation and tubing changes) to de-termine the total cost of therapy. Thecost of sedation-induced prolongation ofventilation or length of ICU stay is likelyto reduce the potential difference in ac-quisition costs between benzodiazepinesand propofol (164). Institutional variables

(bed availability) or patient-specific fac-tors (other injuries and the need for ob-servation) may impact a patient’s lengthof stay more than the sedation regimen.More rapid extubation with propofol wasnot associated with a shorter length ofstay in a multi-center Canadian trial(157). Research that considers all of thesecost factors is needed to estimate theoverall cost of sedative regimens. Sincethe frequency of sedation-induced ad-verse effects has not been well described,there are insufficient data to create phar-macoeconomic models comparing thepotential costs of sedative regimens. Theacquisition costs of sedatives vary widelyamong institutions, and costs may de-cline with the availability of generic prod-ucts (Table 3).

Multidisciplinary development andimplementation of sedation guidelineshave been shown to reduce direct drugcosts (from $81.54 to $18.12 per patientper day), ventilator time (317 to 167hours), and the lengths of ICU stay (19.1to 9.9 days) and total stay without achange in mortality (165). Although aneconomic analysis was not performed, anursing-implemented sedation protocolusing lorazepam reduced the duration ofsedation and mechanical ventilation andthe tracheostomy rate (80). A systematicmultidisciplinary team approach to seda-tion and analgesia will produce clinicaland economic benefits.

An algorithm was developed to incor-porate many of the assessment issueswith the therapy options in this docu-ment (Figure 1). When using this algo-rithm, the pharmacology, potential ad-

Table 5. Clinical trials with one to three days of sedationa

Level ofEvidence

(Reference)Population

(No. Patients) Exclusion Criteria Design Drugs Mean Dosage

1 (158) MICU with respiratoryfailure (73)

. . . Open-label PropofolMidazolamMorphine: p.r.n.

Propofol: 1.25 � 0.87 mg/kg/hrMidazolam: 3.1 � 3.2 mg/hr

2 (118) Surgical or TraumaICU (31)

Alcohol abuse, headinjury, dialysis

Open-label PropofolMidazolamLorazepamMorphine: p.r.n.

Propofol: 2 � 1.5 mg/kg/hrMidazolam: 0.04 � 0.03

mg/kg/hrLorazepam: 0.02 � 0.01

mg/kg/hr2 (157) MICU or SICU (99) Neurosurgery, coma,

seizuresMulticenter, Open-label,

intention-to-treatanalysis



aMICU � medical intensive care unit, ICU � intensive care unit, SICU � surgical intensive care unit.


verse effects, and therapeutic issuesdiscussed in this document should beconsidered.

Recommendations: Midazolam or di-azepam should be used for rapid seda-tion of acutely agitated patients.(Grade of recommendation � C)

Propofol is the preferred sedative whenrapid awakening (e.g., for neurologicassessment or extubation) is impor-tant. (Grade of recommendation � B)

Midazolam is recommended for short-term use only, as it produces unpre-dictable awakening and time to extu-bation when infusions continue longerthan 48–72 hours. (Grade of recom-mendation � A)

Lorazepam is recommended for the se-dation of most patients via intermit-tent i.v. administration or continuousinfusion. (Grade of recommenda-tion � B)

The titration of the sedative dose to adefined endpoint is recommended withsystematic tapering of the dose ordaily interruption with retitration tominimize prolonged sedative effects.(Grade of recommendation � A)

Triglyceride concentrations should bemonitored after two days of propofolinfusion, and total caloric intake fromlipids should be included in the nutri-tion support prescription. (Grade ofrecommendation � B)

The use of sedation guidelines, an algo-rithm, or a protocol is recommended.(Grade of recommendation � B)

SEDATIVE AND ANALGESICWITHDRAWAL

Patients exposed to more than oneweek of high-dose opioid or sedative ther-apy may develop neuroadaptation orphysiological dependence. Rapid discon-tinuation of these agents could lead towithdrawal symptoms. Opioid withdrawalsigns and symptoms include dilation ofthe pupils, sweating, lacrimation, rhinor-rhea, piloerection, tachycardia, vomiting,diarrhea, hypertension, yawning, fever,tachypnea, restlessness, irritability, in-creased sensitivity to pain, cramps, mus-cle aches, and anxiety. Benzodiazepinewithdrawal signs and symptoms includedysphoria, tremor, headache, nausea,sweating, fatigue, anxiety, agitation, in-creased sensitivity to light and sound,paresthesias, muscle cramps, myoclonus,sleep disturbances, delirium, and sei-zures. Propofol withdrawal has not beenwell described but appears to resemblebenzodiazepine withdrawal.

The occurrence of sedative and analge-sic withdrawal has been described in bothadult and pediatric ICU populations (166–168). In adults, withdrawal is associatedwith the length of stay, mechanical venti-lation, and the dose and duration of anal-gesic and sedative therapy. Patients at high-est risk include those who stay greater thanseven days in the ICU, receive greater than35 mg/day of lorazepam, or greater than 5mg/day of fentanyl (166).

Studies of pediatric patients havefound that the rate of medication wean-ing may be very important in the devel-opment of withdrawal syndromes (167,

169). Although not tested prospectively,it has been recommended that daily dosedecrements of opioids not exceed 5–10%in high-risk patients (170). If the drug isadministered intermittently, changingthe therapy to longer-acting agents mayalso attenuate withdrawal symptoms(171). Another recommendation foropioid weaning is to decrease a contin-uous infusion rate by 20 – 40% initiallyand make additional reductions of 10%every 12–24 hours, depending on thepatient’s response (172). Conversion toa continuous subcutaneous infusionhas also been used for gradual fentanyland midazolam weaning in children(173). Patient care costs may be in-creased unnecessarily if sedatives andanalgesics are withdrawn too slowly.

Recommendation: The potential foropioid, benzodiazepine, and propofolwithdrawal should be considered afterhigh doses or more than approxi-mately seven days of continuous ther-apy. Doses should be tapered system-atically to prevent withdrawalsymptoms. (Grade of recommenda-tion � B)

DELIRIUM

As many as 80% of ICU patients havedelirium, characterized by an acutelychanging or fluctuating mental status,inattention, disorganized thinking, andan altered level of consciousness that mayor may not be accompanied by agitation.Placing severely ill patients in a stressfulenvironment for prolonged periods exacer-


Actual Duration Sedation Endpoint/Goal Outcome Measure Significance/Conclusion

Up to 3 days Behavior scale Awakening: defined by eyeopening ability to follow witheyes, hand grasp, and tongueprotrusion

Many awake during infusion,Propofol: smaller range ofawakening times, Used daily wake-up to reassess patients

Propofol: 86.4 � 72 hrMidazolam: 60 � 72 hrLorazepam: 72 � 52.8 hr

Ramsay level 2–4 Time with adequate sedation:Midazolam: 79%Propofol: 62%Lorazepam: 68%

Lorazepam vs. midazolam p � 0.03Midazolam vs. Propofol p � 0.01Propofol: more undersedation 31%Lorazepam: more oversedation14% and precipitation 18%

24–72 hr � post hoc stratumPropofol: n � 21Midazolam: n � 17


Time to extubationPropofol: 7.4 hrMidazolam: 31.3 hr

p � 0.068, Enrollment ended early,insufficient power Overall:Propofol: 60.2% of time at targetRamsay scoreMidazolam: 44% of time at targetRamsay score, p � 0.05


bates the clinical symptoms of delirium(174–177). Delirium is usually character-ized by fluctuating levels of arousalthroughout the day, associated with sleep-wake cycle disruption, and hastened byreversed day-night cycles (178). Delir-ium may be associated with confusionand different motoric subtypes: hypoac-tive, hyperactive, or mixed (179, 180).Hypoactive delirium, which is associ-ated with the worst prognosis, is char-acterized by psychomotor retardationmanifested by a calm appearance, inat-tention, decreased mobility, and obtun-dation in extreme cases. Hyperactivedelirium is easily recognized by agita-tion, combative behaviors, lack of ori-

entation, and progressive confusion af-ter sedative therapy.

Assessment of Delirium

The gold standard criteria used to diag-nose delirium is the clinical history andexamination as guided by the Diagnosticand Statistical Manual of Mental Disor-ders, 4th edition (DSM-IV) (177). Althoughmany scales and diagnostic instrumentshave been developed to facilitate the recog-nition and diagnosis of delirium, thesescales routinely exclude ICU patients be-cause it is often difficult to communicatewith them (178, 181). Several groups ofdelirium investigators have recently collab-

orated to develop and validate a rapid bed-side instrument to accurately diagnose de-lirium in ICU patients, who are oftennonverbal because they are on mechanicalventilation. This instrument is called theConfusion Assessment Method for theICU (CAM-ICU) (181, 182). The workwas begun by Hart and colleagues withtheir publication of the Cognitive Testfor Delirium and a later version calledthe Abbreviated Cognitive Test for De-lirium (183, 184). These two investiga-tions were limited because they in-cluded approximately 20 patients eachand excluded some of the most severelyill patients who are often cared for inthe ICU. These factors led the authors

Table 6. Clinical trials of more than three days of sedation

Level ofEvidence

(Reference)Population

(No. Patients) Exclusion Criteria Design Drugs Mean Dosage

2 (128) MICU/SICUa

(88 total, 28medium-term,20 long-termsedation)

Neurologic injury,ongoing NMBA use



1 (159) MICU (20) CNS abnormal Open-label LorazepamMidazolamMorphine: to all patients

Lorazepam: 0.06 � 0.04 mg/kg/hrMidazolam: 0.24 � 0.16 mg/kg/hr

1 (160) MICU/SICU (98) Renal, hepatic, orcardiac failure,ongoing NMBA use

Open-label,multicenter

PropofolMidazolamMorphine: p.r.n.


1 (129) MICU/SICU (108consecutive)

Chronic liver disease,head injury,ongoing NMBA use


Propofol: 3.07–5.7 0.04 mg/kg/hrMidazolam: 14 � 0.1 mg/kg/hr

2 (70) MICU/SICU (68consecutive)

None Open-label PropofolMidazolamMorphine: p.r.n.


2 (161) MICU/SICU (26) Hepatic or renalinsufficiency, headinjury, ongoingNMBA use

Open-label Propofol � alfentanilMidazolam � morphine

Alfentanil: 0.5–2 �g/kg/hrPropofol: 1–4 mg/kg/hrMorphine: 17–70 �g/kg/hrMidazolam: 0.03–0.2 mg/kg/hr

1 (127) Trauma ICU (100consecutive)

Renal or hepaticfailure

Open-label PropofolMidazolamPropofol � midazolamMorphine: p.r.n.

Propofol: 2.12 � 1.2 mg/kg/hrMidazolam: 0.19 � 0.09 mg/kg/hrPropofol and Midazolam:Propofol: 1.6 � 0.05 mg/kg/hr,Midazolam: 0.14 � 0.08 mg/kg/hr

1 (162) MICU (64) Head injury, ongoingNMBA use

Blinded LorazepamMidazolamFentanyl p.r.n.

Lorazepam: 23.1 � 14.4 mg/dayMidazolam: 372 � 256 mg/day

2 (163) Trauma ICU (63consecutive)

Renal or hepaticfailure

Open-label Propofol 2%MidazolamMorphine: to all patients

Propofol: 6400 � 1797 mg/dayMidazolam: 297.8 � 103.8 mg/day

2 (157) MICU/SICU (99) Neurosurgery, coma,seizures

Multicenter,open-label,intention-to-treatanalysis



aMICU � medical intensive care unit, ICU � intensive care unit, NMBA � neuromuscular blocking agent, GCS � Glasgow Coma Score; SICU, surgicalintensive care unit.


to recommend that additional researchwith delirium assessment tools be con-ducted before routine application inmechanically ventilated patients (184).

Collaboration among specialists in pul-monary and critical care, neurology, psy-chiatry, neuropsychology, and geriatricshas led to the development of a useful as-sessment tool (182, 185). It is based on theConfusion Assessment Method (CAM),which was designed specifically for use byhealth care professionals without formalpsychiatric training, and incorporatesDSM-IV criteria for the diagnosis of delir-ium (186). CAM, which is the most widelyused delirium assessment instrument fornon-psychiatrists, is easy to use and has

demonstrated utility in important clinicalinvestigations (187, 188).

Critical care nurses can complete de-lirium assessments with the CAM-ICU inan average of 2 minutes with an accuracyof 98%, compared with a full DSM-IVassessment by a geriatric psychiatric ex-pert, which usually requires at least 30minutes to complete. The CAM-ICU as-sessments have a likelihood ratio of over50 for diagnosing delirium and high interrater reliability (kappa � 0.96) (185). Inthe two subgroups expected to presentthe greatest challenge to the CAM-ICU(i.e., those over 65 years and those withsuspected dementia), the instrument re-tained excellent inter rater reliability,

sensitivity, and specificity. To completethe CAM-ICU, patients are observed forthe presence of an acute onset of mentalstatus change or a fluctuating mental sta-tus, inattention, disorganized thinking,or an altered level of consciousness (Ta-ble 7). With the CAM-ICU, delirium wasdiagnosed in 87% of the ICU patientswith an average onset on the second dayand a mean duration of 4.2 � 1.7 days(185). Ongoing research will assist in un-derstanding the etiology of delirium andthe effects of therapeutic interventions.

Another instrument for deliriumscreening was validated in ICU patientsby comparison with a psychiatric evalua-tion (189). The use of these tools in pro-


Actual Duration Sedation Endpoint/Goal Outcome Measure Significance/Conclusion

Medium: Propofol: 116 hrMidazolam: 113 hrLong: Propofol: 312 hrMidazolam: 342 hr

Ramsay level 2–5 and modifiedGCS

Time to extubation:Medium: Propofol: 0.4 � 0.1 hrMidazolam: 13.5 � 4 hrLong: Propofol: 0.8 � 0.3 hrMidazolam: 36.6 � 6.8 hr

Medium: p � 0.05 Long: p � 0.05

Lorazepam: 77 hrMidazolam: 108 hr

Ramsay level 2–3 Return to baseline mental status:Lorazepam: 261 � 189 min,Midazolam: 1815 � 2322 min


Propofol: 81 hr Midazolam:88 hr

Own sedation scale Awakening lightly sedated:Propofol: 14 � 0.8 minMidazolam: 64 � 20 minDeep sedation: Propofol: 27 � 16 minMidazolam: 237 � 222 min

Light: p � 0.05 Heavy: p � 0.01


Ramsay level 4–5 Time to t-tube Propofol: 4 � 3.9 hrMidazolam: 48.9 � 47.2 hr

p � 0.0001 Propofol: hightriglycerides 12% men, 50%women


Ramsay level 2–3 also ratedamnesia

Awakening: Propofol: 1.8 � 0.4 hrMidazolam: 2.8 � 0.4 hrAmnesia: Propofol: 29%Midazolam: 100%

p � 0.02 More propofol patientswith agitation upon awakening

. . . Own scale, goal: moderate-heavy sedation

Time to extubation: Propofol: 3 hr(1–13 hr)Midazolam: 50 hr (1–121 hr)

p � 0.006

Propofol: 5.2 daysMidazolam: 6.6 daysPropofol and Midazolam:7.2 days

Own scale, goal: moderate toheavy sedation

Awakening (excluding head trauma):Propofol: 110 � 50 minMidazolam: 660 � 400 minPropofol and Midazolam: 190 � 200 min

Midazolam vs Propofol or Propofoland Midazolam: p � 0.01;Propofol: high triglyceride levels

Lorazepam: 141 hrMidazolam: 141 hr

Addenbrooke sedation scale,initial moderate to heavysedation, tapering to lightsedation

Awakening similar, times not reported Satisfactory sedation: Lorazepam:87 � 10.5%, Midazolam: 66.2 �23.1% p � 0.0001


Own scale, goal: moderate toheavy sedation

Awakening (no head trauma)Propofol: 145 � 50 minMidazolam: 372 � 491 min

Awakening: not statisticallysignificant Less triglycerideelevation than historical control(reference 117)

24–72 hr—post hoc stratumPropofol: n � 4Midazolam: n � 10


Time to extubation:Propofol: 8.4 hrMidazolam: 46.8 hr

p � 0.03, enrollment ended early,insufficient power limitsconclusion Overall: Propofol:60.2% of time at target Ramsayscore Midazolam: 44% of time attarget Ramsay score, p � 0.05


spective trials will delineate the long-term ramifications of delirium on theclinical outcomes of ICU patients. Thestudy of delirium and other forms of cog-nitive impairment in mechanically venti-lated patients and other risk factors forneuropsychological sequelae after ICUcare may be an important advancementin the monitoring and treatment of crit-ically ill patients.

Recommendation: Routine assessmentfor the presence of delirium is recom-mended. (The CAM-ICU is a promisingtool for the assessment of delirium inICU patients.) (Grade of recommenda-tion � B)

Treatment of Delirium

Inappropriate drug regimens for seda-tion or analgesia may exacerbate deliriumsymptoms. Psychotic or delirious pa-tients may become more obtunded andconfused when treated with sedatives,

causing a paradoxical increase in agita-tion (190).

Neuroleptic agents (chlorpromazineand haloperidol) are the most commondrugs used to treat patients with delir-ium. They are thought to exert a stabiliz-ing effect on cerebral function by antag-onizing dopamine-mediated neuro-transmission at the cerebral synapses andbasal ganglia. This effect can also en-hance extrapyramidal symptoms (EPS).Abnormal symptomatology, such as hal-lucinations, delusions, and unstructuredthought patterns, is inhibited, but thepatient’s interest in the environment isdiminished, producing a characteristicflat cerebral affect. These agents also ex-ert a sedative effect.

Chlorpromazine is not routinely usedin critically ill patients because of itsstrong anticholinergic, sedative, and�-adrenergic antagonist effects. Haloper-idol has a lesser sedative effect and alower risk of inducing hypotension than

chlorpromazine. Droperidol, a chemicalcongener of haloperidol, is reported to bemore potent than haloperidol but hasbeen associated with frightening dreamsand may have a higher risk of inducinghypotension because of its direct vasodi-lating and antiadrenergic effects (191,192). Droperidol has not been studied inICU patients as extensively as haloperidol.

Haloperidol is commonly given via in-termittent i.v. injection (193). The opti-mal dose and regimen of haloperidol havenot been well defined. Haloperidol has along half-life (18–54 hours) and loadingregimens are used to achieve a rapid re-sponse in acutely delirious patients. Aloading regimen starting with a 2-mgdose, followed by repeated doses (doublethe previous dose) every 15–20 minuteswhile agitation persists, has been de-scribed (193, 194). High doses of haloper-idol (�400 mg per day) have been re-ported, but QT prolongation may result.However, the safety of this regimen has

Table 7. The confusion assessment method for the diagnosis of delirium in the ICU (CAM-ICU) (182, 185)

Feature Assessment Variables

1. Acute Onset of mental status changes orFluctuating Course

Is there evidence of an acute change in mental status from the baseline?Did the (abnormal) behavior fluctuate during the past 24 hours, i.e., tend to come and go or

increase and decrease in severity?Did the sedation scale (e.g., SAS or MAAS) or coma scale (GCS) fluctuate in the past 24 hours?a

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 (ASE)? (i.e., Visual

Component ASE tests the patient’s ability to pay attention via recall of 10 pictures; auditorycomponent ASE tests attention via having patient squeeze hands or nod whenever the letter“A” is called in a random letter sequence)

3. Disorganized thinking If the patient is already extubated from the ventilator, determine whether or not the patient’sthinking is disorganized or incoherent, such as rambling or irrelevant conversation, unclearor illogical flow of ideas, or unpredictable switching from subject to subject.

For those still on the ventilator, can the patient answer the following 4 questions correctly?1. Will a stone float on water?2. Are there fish in the sea?3. Does one pound weigh more than two pounds?4. Can you use a hammer to pound a nail?

Was the patient able to follow questions and commands throughout the assessment?1. “Are you having any unclear thinking?”2. “Hold up this many fingers.” (examiner holds two fingers in front of patient)?3. “Now do the same thing with the other hand.” (not repeating the number of fingers)

4. Altered level of consciousness (any level ofconsciousness 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 andappropriately interactive when prodded minimally

Stupor: difficult to arouse, unaware of some or all elements in the environment, or notspontaneously interacting with the interviewer; becomes incompletely aware andinappropriately interactive when prodded strongly; can be aroused only by vigorous andrepeated stimuli and as soon as the stimulus ceases, stuporous subjects lapse back into theunresponsive state.

Coma: unarousable, unaware of all elements in the environment, with no spontaneousinteraction or awareness of the interviewer, so that the interview is impossible even withmaximal prodding

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

aSAS � Sedation-Analgesia Scale, MAAS � Motor Activity Assessment Scale, GCS � Glasgow Coma Scale.


been questioned (192, 195–200). Once thedelirium is controlled, regularly scheduleddoses (every four to six hours) may be con-tinued for a few days; then therapy shouldbe tapered over several days. A continuousinfusion of haloperidol (3–25 mg/hr) hasbeen used to achieve more consistent se-rum concentrations (81, 201). The pharma-cokinetics of haloperidol may be affected byother drugs (202).

Neuroleptic agents can cause a dose-dependent QT-interval prolongation of theelectrocardiogram, leading to an increasedrisk of ventricular dysrhythmias, includingtorsades de pointes (195–200). SignificantQT prolongation has been reported withcumulative haloperidol doses as low as 35mg, and dysrhythmias have been reportedwithin minutes of administering i.v. dosesof 20 mg or more (199). A history of cardiacdisease appears to predispose patients tothis adverse event (200). The actual inci-dence of torsades de pointes associated withhalperidol use is unknown, although a his-torical case-controlled study suggests itmay be 3.6% (199).

EPS can occur with these agents. Aslowly eliminated active metabolite ofhaloperidol appears to cause EPS (203–208). EPS has been reported less fre-quently after i.v. versus oral haloperidoladministration, but concurrent benzodi-azepine use may mask EPS appearance.Self-limited movement disorders can beseen several days after tapering or discon-tinuing haloperidol and may last for up totwo weeks (209). Treatment of EPS in-cludes discontinuing the neurolepticagent and a clinical trial of diphenhydra-mine or benztropine mesylate.

Other adverse effects have also beendescribed. Haloperidol therapy for thecontrol of agitation after a traumaticbrain injury may prolong the duration ofposttraumatic amnesia, but the effect onfunctional recovery has not been welldemonstrated in humans (210). Althoughhaloperidol is the most common antipsy-chotic agent associated with neurolepticmalignant syndrome and has been impli-cated in approximately 50% of reportedepisodes (only three cases were reportedin critically ill patients receiving intrave-nous haloperidol), its adverse effects maybe underreported (211–214).

Haloperidol therapy for acutely agi-tated or delirious patients has not beenstudied prospectively in agitated ICU pa-tients, but its utility has been suggestedin case series (81, 193, 201).

Recommendations: Haloperidol is thepreferred agent for the treatment ofdelirium in critically ill patients.(Grade of recommendation � C)

Patients should be monitored for elec-trocardiographic changes (QT intervalprolongation and arrhythmias) whenreceiving haloperidol. (Grade of rec-ommendation � B)

SLEEP

Sleep is believed to be important to re-cover from an illness. Sleep deprivationmay impair tissue repair and overall cellu-lar immune function (215). Sleeplessnessinduces additional stress in critical care pa-tients (3, 216). Allowing a patient to obtainan adequate amount of sleep may be diffi-cult in a critical care unit. Sleep in the ICUhas been characterized by few completesleep cycles, numerous awakenings, and in-frequent rapid-eye-movement (REM) sleep(217). Atypical sleep patterns were demon-strated in critically ill patients receivinghigh doses of sedatives (218).

Sleep Assessment

Similar to pain assessment, the pa-tient’s own report is the best measure ofsleep adequacy, since polysomnographyis not a clinically feasible tool in the crit-ical care setting. If self-report is not pos-sible, systematic observation of a pa-tient’s sleep time by nurses has beenshown to be a valid measure (219). A VASor questionnaire can be used to assesssleep for specific patients (220).

Nonpharmacologic Strategies

Titrating Environmental Stimulation.Nonpharmacologic interventions to pro-mote sleep and increase overall patientcomfort may include environment modi-fication, relaxation, back massage, andmusic therapy. Noise in critical care set-tings is an environmental hazard that dis-rupts sleep (221–223). Sources of noiseinclude equipment, alarms, telephones,ventilators, and staff conversations. Noiselevels above 80 decibels cause arousalfrom sleep. Sleep occurs best below 35decibels (222). Earplugs effectively de-creased noise and increased REM sleep inhealthy volunteers in a study that simu-lated noise heard in an ICU (224). A cre-ative unit design with single rooms mayameliorate noise and provide lightingthat better reflects a day-night orienta-tion (225). Lighting mimicking the 24-

hour day helps patients achieve normalsleep patterns, so bright lights should beavoided at night. In addition, care shouldbe coordinated to minimize frequent in-terruptions during the night.

Relaxation. Head-to-toe relaxation maybenefit anxious critically ill patients whocan follow directions. Relaxation will leadto a parasympathetic response and a de-crease in respiratory rate, heart rate, jawtension, and blood pressure. Relaxationtechniques include deep breathing followedby the sequential relaxation of each musclegroup (226). Relaxation, in combinationwith music therapy, is effective in patientswith myocardial infarction (227).

Music Therapy. Music therapy relaxespatients and decreases their pain. Musicintervention with cardiac surgery patients,during the first postoperative day, de-creased noise annoyance, heart rate, andsystolic blood pressure (228). In mechani-cally ventilated patients, music therapy de-creased anxiety and promoted relaxation(229). Music therapy is a proven interven-tion for anxious patients in other criticalcare settings (229–231). Music can de-crease heart rate, respiratory rate, myocar-dial oxygen demand, and anxiety scores andimprove sleep (232, 233). When selectingmusic, a patient’s personal preferenceshould be considered.

Massage. Back massage is an alterna-tive or adjunct to pharmacologic therapyin critically ill patients. Approximately5–10 minutes of massage initiates therelaxation response and increases a pa-tient’s amount of sleep (234, 235).

Pharmacologic Therapy toPromote Sleep

Patients may remain sleep-deprived de-spite nonpharmacologic interventions.Most patients need a combination of anal-gesics, sedatives, and relaxation techniquesto decrease pain and anxiety and promotesleep. Sedative-hypnotics can induce sleepin healthy individuals, but little is known oftheir use in the critically ill. There was nodifference in sleep quality between twogroups of nonintubated ICU patients re-ceiving midazolam or propofol (59). Oralhypnotics, such as benzodiazepines or zol-pidem, are used in nonintubated patients todecrease sleep latency while increasing to-tal sleep time without affecting sleep archi-tecture in stages three and four and REMsleep (215).

Recommendation: Sleep promotionshould include optimization of the en-


vironment and nonpharmacologicmethods to promote relaxation withadjunctive use of hypnotics. (Grade ofrecommendation � B)

ACKNOWLEDGMENTS

The following professional organizationsand persons reviewed this guideline: theAmerican College of Chest Physicians, theAmerican Academy of Neurology, the Ameri-can Association of Critical Care Nurses, theAmerican Nurses Association, the AmericanPharmaceutical Association, the AmericanCollege of Clinical Pharmacy, Stephanie E.Cooke, Pharm.D., Joe Dasta, Pharm.D., DougFish, Pharm.D., Erkan Hassan, Pharm.D., H.Mathilda Horst, M.D., Kathleen Kelly, M.D.,Karen Kaiser, R.N., Corlayne E. Jackson, M.D.,Maria Rudis, Pharm.D., Carl Schoenberger,M.D., Lori Schoonover, R.N., Gayle Takaniski,Pharm.D., Dan Teres, M.D., FCCM, and KarenThompson, M.D.

REFERENCES

1. 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

2. Society of Critical Care Medicine, AmericanSociety of Health-System Pharmacists: Sed-ative, analgesia, and neuromuscular block-ade of the critically ill adult: Revised clinicalpractice guidelines for 2002. Crit Care Med2002; 30:117–118

3. Novaes MA, Knobel E, Bork AM, et al: Stres-sors in ICU: Perception of the patient, rela-tives and healthcare team. Intensive CareMed 1999; 25:1421–1426

4. Desbiens NA, Wu AW, Broste SK, et al: Painand satisfaction with pain control in seri-ously ill hospitalized adults: Findings fromthe SUPPORT research investigators. CritCare Med 1996; 24:1953–1961

5. Epstein J, Breslow MJ: The stress response ofcritical illness. Crit Care Clin 1999; 15:17–33

6. Lewis KS, Whipple JK, Michael KA, et al:Effect of analgesic treatment on the physi-ological consequences of acute pain. Am JHosp Pharm 1994; 51:1539–1554

7. Mangano DT, Silician D, Hollenberg M, et al:Postoperative myocardial ischemia: Thera-peutic trials using intensive analgesia follow-ing surgery. Anesthesiology 1992; 76:342–353

8. Parker SD, Breslow MJ, Frank SM, et al:Catecholamine and cortisol responses tolower extremity revascularization: Correla-tion with outcome variables. Crit Care Med1995; 23:1954–1961

9. Desai PM: Pain management and pulmonarydysfunction. Crit Care Clin 1999; 15:151–166

10. Gust R, Pecher S, Gust A, et al: Effect ofpatient-controlled analgesia on pulmonarycomplications after coronary artery-bypassgrafting. Crit Care Med 1999; 27:2218–2223

11. Carroll KC, Atkins PJ, Herold GR, et al: Painassessment and management in critically illpostoperative and trauma patients: A mul-tisite study. Am J Crit Care 1999; 8:105–117

12. Ferguson J, Gilroy D, Puntillo K: Dimen-sions of pain and analgesic administrationassociated with coronary artery bypass in anAustralian intensive care unit. J Adv Nurs1997; 26:1065–1072

13. Whipple JK, Lewis KS, Quebbeman EJ, et al:Analysis of pain management in critically illpatients. Pharmacotherapy 1995; 15:592–599

14. Sun X, Weissman C: The use of analgesicsand sedatives in critically ill patients: Phy-sicians’ orders versus medications adminis-tered. Heart Lung 1994; 23:169–176

15. Caswell DR, Williams JP, Vallejo M, et al:Improving pain management in criticalcare. Jt Comm J Qual Improvement 1996;22:702–712

16. Hamill-Ruth RJ, Marohn ML: Evaluation ofpain in the critically ill patient. Crit CareClin 1999; 15:35–54

17. Acute Pain Management Guideline Panel.Acute pain management: Operative or med-ical procedures and trauma. Clinical prac-tice guideline. Rockville, MD: Agency forHealth Care Policy and Research, 1992;AHCPR publication no. 92-0032

18. Ho K, Spence J, Murphy MF: Review of painmanagement tools. Ann Emerg Med 1996;27:427–432

19. Terai T, Yukioka H, Asada A: Pain evalua-tion in the intensive care unit: Observer-reported faces scale compared with self-reported visual analog scale. Reg AnesthPain Med 1998; 23:147–151

20. Puntillo KA: Dimensions of procedural painand its analgesic management in criticallyill surgical patients. Am J Crit Care 1994;3:116–128

21. Meehan DA, McRae ME, Rourke DA, et al:Analgesia administration, pain intensity,and patient satisfaction in cardiac surgicalpatients. Am J Crit Care 1995; 4:435–442

22. Melzack R, Katz J: Pain measurement inpersons in pain. In: Wall PD, Melzack R,eds. Textbook of pain. Edinburgh, Scotland:Churchill Livingstone; 1994:337–351

23. Mateo OM, Krenzischek DA: A pilot study toassess the relationship between behavioralmanifestations of pain and self-report ofpain in post anesthesia care unit patients.J Post Anesth Nurs 1992; 7:15–21

24. Puntillo KA, Miaskowski C, Kehrle K, et al:Relationship between behavioral and physio-logical indicators of pain, critical care pa-tients’ self-reports of pain, and opioid admin-istration. Crit Care Med 1997; 25:1159–1166

25. Sjostrom B, Haljamae H, Dahlgren LO, etal: Assessment of postoperative pain: Impactof clinical experience and professional role.Acta Anaesthesiol Scand 1997; 41:339–344

26. Hall-Lord ML, Larsson G, Steen B: Pain anddistress among elderly intensive care unitpatients: Comparison of patients’ experi-ences and nurses’ assessments. Heart Lung1998; 27:123–132

27. Stannard D, Puntillo K, Miaskowski C, et al:Clinical judgement and management ofpostoperative pain in critical care patients.Am J Crit Care 1996; 5:433–441

28. Desbiens NA, Mueller-Rizner N: How welldo surrogates assess the pain of seriouslyill patients? Crit Care Med 2000; 28:1347–1352

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

30. Wagner BKJ, O’Hara DA: Pharmacokineticsand pharmacodynamics of sedatives and an-algesics in the treatment of agitated criti-cally ill patients. Clin Pharmacokinet 1997;33:426–453

31. Barr JB, Donner A: Optimal intravenous dos-ing strategies for sedatives and analgesics inthe ICU. Crit Care Clin 1995; 11:827–847

32. AmeriSource Corporation, ECHO software,version 3.1Q, release 00112. Accessed 2001Feb 6

33. Danziger LH, Martin SJ, Blum RA: Centralnervous system toxicity associated with me-peridine use in hepatic disease. Pharmaco-therapy 1994; 14:235–238

34. Hangmeyer KO, Mauro LS, Mauro VF: Me-peridine-related seizures associated withpatient-controlled analgesia pumps. AnnPharmacother 1993; 27:29–32

35. Tipps LB, Coplin WM, Murry KR, et al: Safetyand feasibility of continuous infusion ofremifentanil in the neurosurgical intensivecare unit. Neurosurgery 2000; 46:596–602

36. Lay TD, Puntillo KA, Miaskowski MI: Anal-gesics prescribed and administered to in-tensive care cardiac surgery patients: Doespatient age make a difference? Prog Cardio-vasc Nurs 1996; 11:17–24

37. Macintyre PE, Jarvis DA: Age is the bestpredictor of postoperative morphine re-quirements. Pain 1995; 64:357–364

38. Fraser GL, Prato S, Berthiaume D, et al:Evaluation of agitation in ICU patients: In-cidence, severity, and treatment in theyoung versus the elderly. Pharmacotherapy2000; 20:75–82

39. Hofbauer R, Tesinsky P, Hammerschmidt V,et al: No reduction in the sufentanil require-ment of elderly patients undergoing ventila-tory support in the medical intensive careunit. Eur J Anaesthesiol 1999; 16:702–707

40. McArdle P: Intravenous analgesia. Crit CareClin 1999; 15:89–105

41. Grossman M, Abiose A, Tangphao O, et al:Morphine-induced venodilation in humans.Clin Pharmacol Ther 1996; 60:554–560

42. Flacke JW, Flacke WE, Bloor BC, et al: His-tamine release by four narcotics: A double-blind study in humans. Anesth Analg 1987;66:723–730

43. Yuan C, Foss JF, O’Connor MF, et al: Effectsof low-dose morphine on gastric emptyingin healthy volunteers. J Clin Pharmacol1998; 38:1017–1020

44. Heyland DK, Tougas G, King D, et al: Im-paired gastric emptying in mechanically


ventilated, critically ill patients. IntensiveCare Med 1996; 22:1339–1344

45. Bosscha K, Nieuwenhuijs VB, Vos A, et al:Gastrointestinal motility and gastric tubefeeding in mechanically ventilated patients.Crit Care Med 1998; 26:1510–1517

46. Albanese J, Viviand X, Potie F, et al: Sufen-tanil, fentanyl, and alfentanil in headtrauma patients: A study on cerebral hemo-dynamics. Crit Care Med 1999; 27:407–411

47. Sperry RJ, Bailey PL, Reichman MV, et al:Fentanyl and sufentanil increase intracra-nial pressure in head trauma patients. An-esthesiology 1992; 77:416–420

48. DeNaal M, Munar F, Poca MA, et al: Cere-bral hemodynamic effects of morphine andfentanyl in patients with severe head injury.Anesthesiology 2000; 92:11–19

49. Dasta JF, Fuhrman TM, McCandles C: Pat-terns of prescribing and administeringdrugs for agitation and pain in patients in asurgical intensive care unit. Crit Care Med1994; 22:974–980

50. 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

51. Boldt J, Thaler E, Lehmann A, et al: Painmanagement in cardiac surgery patients:Comparison between standard therapy andpatient-controlled analgesia regimen. J Car-diothorac Vasc Anesth 1998; 12:654–658

52. Pearce CJ, Gonzalez FM, Wallin JD: Renalfailure and hyperkalemia associated withketorolac tromethamine. Arch Intern Med1993; 153:1000–1002

53. Schlondorff D: Renal complications of non-steroidal anti-inflammatory drugs. KidneyInt 1993; 44:643–653

54. Feldman HI, Kinman JL, Berlin JA, et al:Parenteral ketorolac: The risk for acute renalfailure. Ann Intern Med 1997; 126:193–199

55. Strom BL, Berlin JA, Kinman JL, et al:Parenteral ketorolac and risk of gastrointes-tinal and operative site bleeding. JAMA1996; 275:376–382

56. Bombardier C, Laine L, Reicin A, et al:Comparison of upper gastrointestinal toxic-ity of rofecoxib and naproxen in patientswith rheumatoid arthritis. N Engl J Med2000; 343:1520–1528

57. Peduta VA, Ballabio M, Stefanini S: Efficacyof propacetamol in the treatment of postop-erative pain. Morphine sparing effect in or-thopedic surgery. Italian CollaborativeGroup on Propacetamol. Acta AnaesthesiolScand 1998; 42:293–298

58. Zimmerman HJ, Maddrey W: Acetaminophenhepatotoxicity with regular intake of alcohol.Analysis of instances of therapeutic misadven-ture. Hepatology 1995; 22:767–773

59. Treggiari-Venzi M, Borgeat A, Fuchs-BuderT, et al: Overnight sedation with midazolamor propofol in the ICU: Effects on sleepquality, anxiety and depression. IntensiveCare Med 1996; 22:1186–1190

60. Jenkins DH: Substance abuse and with-

drawal in the intensive care unit: Contem-porary issues. Surg Clin North Am 2000;80:1033–1053

61. Cornwell EE, Belzberg H, Velmahos G, et al:The prevalence and effect of alcohol anddrug abuse on cohort-matched critically in-jured patients. Am Surg 1998; 64:461–465

62. Spies CD, Rommelspacher H: Alcohol with-drawal in the surgical patient: Prevention andtreatment. Anesth Analg 1999; 88:946–954

63. Kaye AD, Clarke RC, Sabar R, et al: Herbalmedicines: Current trends in anesthesiol-ogy practice—a hospital survey. J ClinAnesth 2000; 12:468–471

64. Atkins PM, Mion LC, Mendelson W, et al:Characteristics and outcomes of patientswho self-extubate from ventilatory support:A case-control study. Chest 1997; 112:1317–1323

65. Vassal T, Anh NGD, Gabillet JM, et al: Pro-spective evaluation of self-extubations in amedical intensive care unit. Intensive CareMed 1993; 19:340–342

66. Fraser GL, Riker RR, Prato BS, et al: Thefrequency and cost of patient-initiated de-vice removal in the ICU. Pharmacotherapy2001; 21:1–6

67. Conti J, Smith D: Haemodynamic responsesto extubation after cardiac surgery with andwithout continued sedation. Br J Anaesth1998; 80:834–836

68. Cohen D, Horiuchi K, Kemper M, et al:Modulating effects of propofol on metabolicand cardiopulmonary responses to stressfulintensive care unit procedures. Crit CareMed 1996; 24:612–617

69. Veselis RA, Reinsel RA, Feshchenko VA, etal: The comparative amnestic effects of mi-dazolam, propofol, thiopental, and fentanylat equi-sedative concentrations. Anesthesi-ology 1997; 87:749–764

70. Weinbroum AA, Halpern P, Rudick V, et al:Midazolam versus propofol for long-termsedation in the ICU: A randomized prospec-tive comparison. Intensive Care Med 1997;23:1258–1263

71. Wagner BK, O’Hara DA, Hammond JS:Drugs for amnesia in the ICU. Am J CritCare 1997; 6:192–201

72. Wagner BK, Zavotsky KE, Sweeney JB, et al:Patient recall of therapeutic paralysis in asurgical critical care unit. Pharmacother-apy 1998; 18:358–363

73. Schelling G, Stoll C, Meier M, et al: Health-related quality of life and post-traumaticstress disorder in survivors of adult respira-tory distress syndrome. Crit Care Med 1998;26:651–659

74. Jones C, Griffiths RD: Disturbed memoryand amnesia related to intensive care. Mem-ory 2000; 8:79–94

75. Jones C, Griffiths RD, Humphris G, et al:Memory, delusions, and the development ofacute post-traumatic stress disorder-relatedsymptoms after intensive care. Crit CareMed 2001; 29:573–580

76. Schnyder U, Morgeli H, Nigg C, et al: Early

psychological reactions to life-threateninginjuries. Crit Care Med 2000; 28:86–92

77. Scragg P, Jones A, Fauvel N: Psychologicalproblems following ICU treatment. Anaes-thesia 2001; 56:9–14

78. Watling SM, Johnson M, Yanos J: A methodto produce sedation in critically ill patients.Ann Pharmacother 1996; 30:1227–1231

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

80. 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

81. Riker RR, Fraser GL, Cox PM: Continuousinfusion of haloperidol controls agitation incritically ill patients. Crit Care Med 1994;22:89–97

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

83. Devlin JW, Boleski G, Mlynarek M, et al:Motor Activity Assessment Scale: A validand reliable sedation scale for use with me-chanically ventilated patients in an adultsurgical intensive care unit. Crit Care Med1999; 27:1271–1275

84. Ramsay MA, Savege TM, Simpson BR, et al:Controlled sedation with alphaxalone-alphadolone. Br Med J 1974; 2:656–659

85. Hansen-Flaschen J, Cowen J, Polomano 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

86. DeLemos J, Tweeddale M, Chittock DR:Measuring quality of sedation in adult me-chanically ventilated critically ill patients:The Vancouver Interaction and CalmnessScale. J Clin Epidemiol 2000; 53:908–919

87. Chernik DA, Gillings D, Laine H, et al: Va-lidity and reliability of the Observer’s As-sessment of Alertness/Sedation Scale: Studywith intravenous midazolam. J ClinPsychopharmacol 1990; 10:244–251

88. Ambuel B, Hamlett KW, Marx CM, et al:Assessing distress in pediatric intensivecare environments: The COMFORT scale.J Pediatr Psychol 1992; 17:95–109

89. Rosow C, Manberg PJ: Bispectral indexmonitoring. Anesthesiol Clin North Am1998; 2:89–107

90. Liu J, Singh H, White PF: Electroencephalo-graphic bispectral index correlates with intra-operative recall and depth of propofol-inducedsedation. Anesth Analg 1997; 84:185–189

91. Doi M, Gajraj RJ, Mantzaridis H, et al: Re-lationship between calculated blood con-centration of propofol and electrophysiolog-ical variables during emergence fromanesthesia: Comparison of bispectral index,spectral edge frequency, median frequency,and auditory evoked potential index. Br JAnaesth 1997; 78:180–184

92. Liu J, Singh H, White PF: Electroencepha-


logram bispectral analysis predicts thedepth of midazolam-induced sedation. An-esthesiology 1996; 84:64–69

93. Simmons LE, Riker RR, Prato BS, et al:Assessing sedation levels in mechanicallyventilated ICU patients with the bispectralindex and the sedation-agitation scale. CritCare Med 1999; 27:1499–1504

94. Riker RR, Simmons LE, Fraser GL, et al:Validating the sedation-agitation scale withthe bispectral index and visual analog scalein adult ICU patients after cardiac surgery.Intensive Care Med 2001; 27:853–858

95. De Deyne C, Struys M, Decruyenaere J, et al:Use of continuous bispectral EEG monitoringto assess depth of sedation in ICU patients.Intensive Care Med 1998; 24:1294–1298

96. Ghoneim MM, Mewaldt SP: Benzodiaz-epines and human memory: A review.Anesthesiology 1990; 72:926 –938

97. Gilliland HE, Prasad BK, Mirakhur RK, etal: An investigation of the potential mor-phine sparing effect of midazolam. Anaes-thesia 1996; 51:808–811

98. Greenblatt DJ, Ehrenberg BL, GundermanJ, et al: Kinetic and dynamic study of intra-venous lorazepam: Comparison with intra-venous diazepam. J Pharmacol Exp Ther1989; 250:134–139

99. Vree TB, Shimoda M, Driessen JJ, et al:Decreased plasma albumin concentrationresults in increased volume of distributionand decreased elimination of midazolam inintensive care patients. Clin PharmacolTher 1989; 46:537–544

100. Driessen JJ, Vree TB, Guelen PJ: The effectsof acute changes in renal function on thepharmacokinetics of midazolam duringlong-term infusion in ICU patients. ActaAnaesthesiol Belg 1991; 42:149–155

101. Greenblatt DJ, Abernethy DR, Locniskar Aet al. Effect of age, gender, and obesity onmidazolam kinetics. Anesthesiology 1984;61:27–35

102. Patel IH, Soni PP, Fukuda EK, et al: Thepharmacokinetics of midazolam in patientswith congestive heart failure. Br J ClinPharmacol 1990; 29:565–569

103. Macgilchrist AJ, Birnie GG, Cook A, et al:Pharmacokinetics and pharmacodynamics ofintravenous midazolam in patients with se-vere alcoholic cirrhosis. Gut 1986; 27:190–195

104. Boulieu R, Lehmann B, Salord F, et al:Pharmacokinetics of midazolam and itsmain metabolite 1-hydroxymidazolam inintensive care patients. Eur J Drug MetabPharmacokinet 1998; 23:255–258

105. Malacrida R, Fritz ME, Suter P, et al: Phar-macokinetics of midazolam administered bycontinuous infusion to intensive care pa-tients. Crit Care Med 1992; 20:1123

106. Bauer TM, Ritz R, Haberthur C, et al: Pro-longed sedation due to accumulation ofconjugated metabolites of midazolam. Lan-cet 1995; 346:145–147

107. Hamaoka N, Oda Y, Hase I, et al: Propofoldecreases the clearance of midazolam by in-

hibiting CYP3A4: An in vivo and in vitrostudy. Clin Pharmacol Ther 1999; 66:110–117

108. Gorski JC, Jones DR, Haehner-Daniels BD,et al: The contribution of intestinal andhepatic CYP3A to the interaction betweenmidazolam and clarithromycin. Clin Phar-macol Ther 1998; 64:133–143

109. Ahonen J, Olkkola KT, Salmenpera M, et al:Effect of diltiazem on midazolam and alfen-tanil disposition in patients undergoingcoronary artery bypass grafting. Anesthesi-ology 1996; 85:1246–1252

110. Bailie GR, Cockshott ID, Douglas EJ, et al:Pharmacokinetics of propofol during andafter long-term continuous infusion formaintenance of sedation in ICU patients.Br J Anaesth 1992; 68:486–491

111. Hammerlein A, Derendorf H, LowenthalDT: Pharmacokinetic and pharmacody-namic changes in the elderly. Clin Pharma-cokinet 1998; 35:49–64

112. Michalets EL: Update: Clinically significantcytochrome P-450 drug interactions.Pharmacotherapy 1999; 18:84–112

113. Shelly MP, Sultan MA, Bodenham A, et al:Midazolam infusions in critically ill pa-tients. Eur J Anaesthesiol 1991; 8:21–27

114. Nishiyama T: Sedation during artificial ven-tilation by continuous intravenous infusionof midazolam: Effects of hepatocellular orrenal damage. J Intensive Care Med 1997;12:40–44

115. Ariano RE, Kassum DA, Aronson KJ:Comparison of sedative recovery time aftermidazolam versus diazepam administra-tion. Crit Care Med 1994; 22:1492–1496

116. Newton DW, Narducci WA, Leet WA, et al:Lorazepam stability in and sorption fromintravenous admixture solutions. Am JHosp Pharm 1983; 40:424–427

117. Hoey LL, Vance-Bryan K, Clarens DM, et al:Lorazepam stability in parenteral solutionsfor continuous intravenous administration.Ann Pharmacother 1996; 30:343–346

118. McCollam JS, O’Neil MG, Norcross ED, etal: Continuous infusions of lorazepam, mi-dazolam and propofol for sedation of thecritically-ill surgery trauma patient: A pro-spective, randomized comparison. Crit CareMed 1999; 27:2454–2458

119. Laine GA, Hamid Hossain SM, Solis RT, et al:Polyethylene glycol nephrotoxicity secondaryto prolonged high-dose intravenous loraz-epam. Ann Pharmacother 1995; 29:1110–1114

120. Seay RE, Graves PJ, Wilkin MK: Comment:Possible toxicity from propylene glycol inlorazepam infusion. Ann Pharmacother1997; 31:647–648

121. Arbour RB: Propylene glycol toxicity relatedto high-dose lorazepam infusion: Case re-port and discussion. Am J Crit Care 1999;8:499–506

122. Lugo RA, Chester EA, Cash J, et al: A costanalysis of enterally administered loraz-epam in the pediatric intensive care unit.Crit Care Med 1999; 27:417–421

123. Shepherd MF, Felt-Gunderson PA: Diarrheaassociated with lorazepam solution in a tube-fed patient. Nutr Clin Pract 1996; 11:117–120

124. Kamijo Y, Masuda T, Nishikawa T, et al:Cardiovascular response and stress reactionto flumazenil injection in patients underinfusion with midazolam. Crit Care Med2000; 28:318–323

125. Breheny FX: Reversal of midazolam seda-tion with flumazenil. Crit Care Med 1992;20:736–739

126. Kowalski SD, Rayfield CA: A post hoc de-scriptive study of patients receiving propo-fol. Am J Crit Care 1999; 8:507–513

127. 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

128. Carrasco G, Molina R, Costa J, et al: Propo-fol versus midazolam in short-, medium-,and long-term sedation of critically ill pa-tients. Chest 1993; 103:557–564

129. 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; 25:33–40

130. Kumar AN, Achwartz DE, Lim KG: Propo-fol-induced pancreatitis: Recurrence of pan-creatitis after rechallenge. Chest 1999; 115:1198–1199

131. Possidente CJ, Rogers FB, Osler TM, et al:Elevated pancreatic enzymes after extendedpropofol therapy. Pharmacotherapy 1998;18:653–655

132. Leisure GS, O’Flaherty J, Green L, et al:Propofol and postoperative pancreatitis.Anesthesiology 1996; 84:224–227

133. Bray RJ: Propofol infusion syndrome in chil-dren. Paediatr Anaesth 1998; 8:491–499

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

135. 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

136. Webb SA, Roberts B, Breheny FX, et al:Contamination of propofol infusions in theintensive care unit: Incidence and clinicalsignificance. Anaesth Intensive Care 1998;26:162–164

137. Avramov MN, Husain MM, White PF: Thecomparative effects of methohexital, propo-fol, and etomidate for electroconvulsivetherapy. Anesth Analg 1995; 81:596–602

138. Stecker MM, Kramer TH, Raps EC, et al:Treatment of refractory status epilepticuswith propofol: Clinical and pharmacoki-netic findings. Epilepsia 1998; 39:18–26

139. McCowan C, Marik P: Refractory deliriumtremens treated with propofol: A case series.Crit Care Med 2000; 28:1781–1784

140. Herregods L, Verbeke J, Rolly G, et al: Effectof propofol on elevated intracranial pres-


sure. Preliminary results. Anaesthesia1988; 43(suppl):107–109

141. 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

142. Herregods L, Mergaert C, Rolly G, et al: Com-parison of the effects of 24-hour propofol orfentanyl infusions on intracranial pressure. JDrug Dev 1989; 2(suppl 2):99–100

143. Spies CD, Dubisz N, Neumann T, et al:Therapy of alcohol withdrawal syndrome inintensive care unit patients following trau-ma: Results of a prospective, randomizedtrial. Crit Care Med 1996; 24:414–422

144. Ip Yam PC, Forbes A, Kox WJ: Clonidine inthe treatment of alcohol withdrawal in theintensive care unit. Br J Anaesth 1992; 69:328

145. Aantaa RE, Kanto JH, Scheinin M: Dexme-detomidine pre-medication for minor gyneco-logic surgery. Anesth Analg 1990; 4:407–413

146. Venn RM, Bradshaw CJ, Spencer R, et al:Preliminary UK experience of dexmedeto-midine, a novel agent for postoperative se-dation in the intensive care unit. Anaesthe-sia 1999; 54:1136–1142

147. Hall JE, Uhrich TD, Barnet JA, et al: Seda-tive, amnestic, and analgesic properties ofsmall-dose dexmedetomidine infusions.Anesth Analg 2000; 90:699–705

148. Belleville JP, Ward DS, Bloor BC, et al: Effectsof intravenous dexmedetomidine in humans.Anesthesiology 1992; 77:1125–1133

149. Osterman ME, Keenan SP, Seiferling RA, etal: Sedation in the intensive care unit, a sys-tematic review. JAMA 2000; 283:1451–1459

150. Aitkenhead AR, Willatts SM, Park GR, et al:Comparison of propofol and midazolam forsedation in critically ill patients. Lancet1989; 2(8665):704–709

151. Roekaerts PM, Huygen FJ, deLange S: Infu-sion of propofol versus midazolam for seda-tion in the intensive care unit followingcoronary artery surgery. J CardiothoracVasc Anesth 1993; 7:142–147

152. Higgins TL, Yared JP, Estafanous FG, et al:Propofol versus midazolam for intensive careunit sedation after coronary artery bypassgrafting. Crit Care Med 1994; 22:1415–1423

153. 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

154. Cernaianu AC, DelRossi AJ, Flum DR, et al:Lorazepam and midazolam in the intensivecare unit: a randomized, prospective, mul-ticenter study of hemodynamics, oxygentransport, efficacy, and cost. Crit Care Med1996; 24:222–228

155. Searle NR, Taillefer J, Gagnon L, et al:Propofol or midazolam for sedation andearly extubation following cardiac surgery.Can J Anaesth 1997; 44:629–635

156. 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

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

158. Kress JP, O’Connor MF, Pohlman AS, et al:Sedation of critically ill patients during me-chanical ventilation. Am J Resp Crit CareMed 1996; 153:1012–1018

159. 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

160. Chamorro C, DeLatorre FJ, Montero A, et al:Comparative study of propofol versus midazo-lam in the sedation of critically ill patients:Results of a prospective, randomized, multi-center trial. Crit Care Med 1996; 24:932–939

161. Manley NM, Fitzpatrick RW, Long T, et al: Acost analysis of alfentanil � propofol vs.morphine � midazolam for the sedation ofcritically ill patients. Pharmacoeconomics1997; 12:247–255

162. Swart E, van Schijndel RJM, van LoenenAC, et al: Continuous infusion of lorazepamversus midazolam in patients in the inten-sive care unit: Sedation with lorazepam iseasier to manage and is more cost effective.Crit Care Med 1999; 27:1461–1465

163. Sandiumenge Camps A, Sanchez-IzquierdoRiera JA, Toral Vazquez D, et al: Midazolamand 2% propofol in long-term sedation oftraumatized critically ill patients: Efficacyand safety comparison. Crit Care Med 2000;28:3612–3619

164. Kollef MH, Levy NT, Ahrens TS, et al: Theuse of continuous i.v. sedation is associatedwith prolongation of mechanical ventila-tion. Chest 1998; 114:541–548

165. Mascia MF, Koch M, Medicis JJ: Pharmaco-economic impact of rational use guidelineson the provision of analgesia, sedation, andneuromuscular blockade in critical care.Crit Care Med 2000; 28:2300–2306

166. Cammarano WB, Pittet JF, Weitz S, et al:Acute withdrawal syndrome related to theadministration of analgesic and sedativemedications in adult intensive care unit pa-tients. Crit Care Med 1998; 26:676–684

167. Katz R, Kelly HW, His A: Prospective study onthe occurrence of withdrawal in critically illchildren who received fentanyl by continuousinfusion. Crit Care Med 1994; 22:763–777

168. Fonsmark L, Rasmussen YH, Carl P: Occur-rence of withdrawal in critically ill children.Crit Care Med 1999; 27:196–199

169. Carr DB, Todres ID: Fentanyl infusion andweaning in the pediatric intensive care unit:Toward science-based practice. Crit CareMed 1994; 22:725–727

170. Buck M, Blumer J: Opioids and other anal-gesics: Adverse effects in the intensive careunit. Crit Care Clin 1991; 7:615–637

171. American Pain Society. Principles of anal-

gesic use in the treatment of acute pain andcancer pain. 3rd ed. Skokie, IL: 1992.

172. Anand KJS, Ingraham J: Tolerance, depen-dence, and strategies for compassionate with-drawal of analgesics and anxiolytics in thepediatric ICU. Crit Care Nurse 1996; 16:87–93

173. Tobias JD: Subcutaneous administration offentanyl and midazolam to prevent with-drawal after prolonged sedation in children.Crit Care Med 1999; 27:2262–2265

174. Koolhoven I, Tjon-A-Tsien MRS, van derMast RC: Early diagnosis of delirium aftercardiac surgery. Gen Hosp Psychiatry 1996;18:448–451

175. Sanders KM, Stern TA, O’Gara PT, et al:Delirium during intra-aortic balloon pumptherapy. Psychosomatics 1992; 33:35–44

176. Dyer CB, Ashton CM, Teasdale TA: Postop-erative delirium. A review of 80 primarydata collection studies. Arch Intern Med1995; 155:461–465

177. American Psychiatric Association. DSM-IV:Diagnostic and statistical manual of mentaldisorders. Washington, DC: American Psy-chiatric Press; 1994.

178. Trzepacz PT: Delirium—advances in diag-nosis, pathophysiology, and treatment. Psy-chiatr Clin North Am 1996; 19:429–449

179. Lipowski ZJ: Delirium in the elderly patient.N Engl J Med 1989; 320:578–582

180. Meagher DJ, Trzepacz PT: Motoric subtypesof delirium. Semin Clin Neuropsychiatry2000; 5:75–85

181. Ely EW, Siegel MD, Inouye S: Delirium inthe intensive care unit: an under-recog-nized syndrome of organ dysfunction. Se-min Resp Crit Care Med 2001; 22:115–126

182. 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

183. Hart RP, Levenson JL, Sessler CN, et al:Validation of a cognitive test for delirium inmedical ICU patients. Psychosomatics1996; 37:533–546

184. Hart RP, Best ALN, Sessler CN, et al: Ab-breviated cognitive test for delirium. J Psy-chosom Res. 1997; 43:417–423

185. Ely EW, Inouye SK, Bernard GR, et al: De-lirium in mechanically ventilated patients:Validity and reliability of the confusion as-sessment method for the intensive care unit(CAM-ICU). JAMA 2001; 286:2703–2710

186. Inouye SK, van Dyck CH, Alessi CA, et al:Clarifying confusion: the confusion assess-ment method. Ann Intern Med 1990; 113:941–948

187. Smith MJ, Breitbart WS, Meredith MP: Acritique of instruments and methods to de-tect, diagnose, and rate delirium. J PainSymptom Manage 1995; 10:35–77

188. Inouye SK, Bogardus ST, Charpentier PA, etal: A multicomponent intervention to pre-vent delirium in hospitalized older patients.N Engl J Med 1999; 340:669–676

189. Bergeron N, Dubois MJ, Dumont M, et al:Intensive care delirium screening check-


list: Evaluation of a new screening tool.Intensive Care Med 2001; 27:859 – 864

190. Breitbart W, Marotta R, Platt MM, et al: Adouble-blind trial of haloperidol, chlor-promazine, and lorazepam in the treatmentof delirium in hospitalized AIDS patients.Am J Psychiatry 1996; 153:231–237

191. Frye MA, Coudreaut MF, Hakeman SM, et al:Continuous droperidol infusion for manage-ment of agitated delirium in an intensive careunit. Psychosomatics 1995; 36:301–305

192. Faigel DO, Metz DC, Kochman ML: Torsadede pointes complicating the treatment ofbleeding esophageal varices: Associationwith neuroleptics, vasopressin, and electro-lyte imbalance. Am J Gastroenterol 1995;90:822–824

193. Tesar GE, Murray GB, Cassem NH: Use ofhaloperidol for acute delirium in intensivecare setting. J Clin Psychopharmacol 1985;5:344–347

194. Tesar GE, Stern TA: Rapid tranquilization ofthe agitated intensive care unit patient.J Intensive Care Med 1988; 3:195–201

195. DiSalvo TG, O’Gara PT: Torsade de pointescaused by high-dose intravenous haloperi-dol in cardiac patients. Clin Cardiol 1995;28:285–290

196. Metzger E, Friedman R: Prolongation of thecorrected QT and torsades de pointes car-diac arrhythmia associated with intrave-nous haloperidol in the medically ill. J ClinPsychopharmacol 1993; 13:128–132

197. Wilt JL, Minnema AM, Johnson RF, et al:Torsade de pointes associated with the useof intravenous haloperidol. Ann Intern Med1993; 119:391–394

198. Hunt N, Stern TA: The association betweenintravenous haloperidol and torsades depointes. Psychosomatics 1995; 36:541–549

199. Sharma ND, Rosman HS, Padhi D, et al:Torsades de pointes associated with intra-venous haloperidol in critically ill patients.Am J Cardiol 1998; 81:238–240

200. Lawrence KR, Nasraway SA: Conductiondisturbances associated with administrationof butyrophenone antipsychotics in the crit-ically ill: A review of the literature. Phar-macotherapy 1997; 17:531–537

201. Seneff MG, Mathews RA: Use of haloperidolinfusions to control delirium in critically illadults. Ann Pharmacother 1995; 29:690–693

202. Kudo S, Ishizaki T: Pharmacokinetics ofhaloperidol. Clin Pharmacokinet 1999; 37:435–456

203. Fang J, Baker GB, Silverstone PH, et al:Involvement of CYP3A4 and CYP2D6 inthe metabolism of haloperidol. Cell MolNeurobiol 1997; 17:227–233

204. Tsang MW, Shader RI, Greenblatt DJ: Me-tabolism of haloperidol: Clinical implica-tions and unanswered questions. J ClinPsychopharmacol 1994; 14:159–162

205. Avent KM, Riker RR, Fraser GL, et al: Me-tabolism of haloperidol to pyridinium spe-cies in patients receiving high doses intra-

venously: is HPTP an intermediate? Life Sci1997; 61:2383–2390

206. Eyles DW, McLennan HR, Jones A, et al:Quantitative analysis of two pyridinium me-tabolites of haloperidol in patients withschizophrenia. Clin Pharmacol Ther 1994;56:512–520

207. Usuki E, Pearce R, Parkinson A, et al: Stud-ies on the conversion of haloperidol and itstetrahydropyridine dehydration product topotentially neurotoxic pyridinium metabo-lites by human liver microsomes. Chem ResToxicol 1996; 9:800–806

208. Eyles DW, McGrath JJ, Pond SM: Formationof pyridinium species of haloperidol in hu-man liver and brain. Psychopharmacology1996; 125:214–219

209. Riker RR, Fraser GL, Richen P: Movementdisorders associated with withdrawal fromhigh-dose intravenous haloperidol therapyin delirious ICU patients. Chest 1997; 111:1778–1781

210. Rao N, Jellinek HM, Woolston DC: Agitationin closed head injury: Haloperidol effects onrehabilitation outcome. Arch Phys Med Re-habil 1985; 66:30–34

211. Shalev A, Hermesh H, Munitz H: Mortalityfrom neuroleptic malignant syndrome.J Clin Psychiatry 1989; 50:18–25

212. Still J, Friedman B, Law E, et al: Neurolep-tic malignant syndrome in a burn patient.Burns 1998; 24:573–575

213. Levitt AJ, Midha R, Craven JL: Neurolepticmalignant syndrome with intravenous halo-peridol. Can J Psychiatry 1990; 35:789 Letter.

214. Burke C, Fulda GJ, Casellano J: Neurolepticmalignant syndrome in a trauma patient.J Trauma 1995; 39:796–798

215. Krachman SL, D’Alonzo GE, Criner GJ:Sleep in the intensive care unit. Chest 1995;107:1713–1720

216. Stein-Parbury J, McKinley S: Patients’ ex-periences of being in an intensive care unit:A select literature review. Am J Crit Care2000; 9:20–27

217. Redeker NS: Sleep in acute care settings: Anintegrative review. J Nurs Scholarsh 2000;32:31–38

218. Cooper AB, Thornley KS, Young GB, et al:Sleep in critically ill patients requiring me-chanical ventilation. Chest 2000; 117:809–818

219. Edwards GB, Schuring LM: Pilot study: Val-idating staff nurses’ observations of sleepand wake states among critically ill patientsusing polysomnography. Am J Crit Care1993; 2:125–131

220. Richardson SJ: A comparison of tools forthe assessment of sleep pattern disturbancein critically ill adults. Dimens Crit CareNurs 1997; 16:226–239

221. Meyers TJ, Eveloff SE, Bauer MS, et al:Adverse environmental conditions in the re-spiratory and medical ICU settings. Chest1994; 105:1211–1216

222. Aaron JN, Carlisle CC, Carskadon MA, et al:Environmental noise as a cause of sleep

disruption in an intermediate respiratorycare unit. Sleep 1996; 19:707–710

223. Freedman NS, Kotzer N, Schwab RJ: Pa-tient perception of sleep quality and etiol-ogy of sleep disruption in the intensive careunit. Am J Resp Crit Care Med 1999; 159:1155–1162

224. Wallace CJ, Robins J, Alvord LS, et al: Theeffect of earplugs on sleep measures duringexposure to simulated intensive care unitnoise. Am J Crit Care 1999; 8:210–219

225. Horsburgh CR: Healing by design. N EnglJ Med 1995; 333:735–740

226. Griffin JP, Myers S, Kopelke C, et al: Theeffects of progressive muscular relaxationon subjectively reported disturbance due tohospital noise. Behav Med 1988; 14:37–42

227. Guzzetta CE: Effects of relaxation and mu-sic therapy on patients in a coronary careunit with presumptive myocardial infarc-tion. Heart Lung 1989; 18:609–616

228. Byers JF, Smyth KA: Effect of a music in-tervention on noise annoyance, heart rate,and blood pressure in cardiac surgery pa-tients. Am J Crit Care 1997; 6:183–191

229. Chlan L: Effectiveness of a music therapyintervention on relaxation and anxiety forpatients receiving ventilatory assistance.Heart Lung 1998; 27:169–176

230. Updike P: Music therapy results for ICU pa-tients. Dimens Crit Care Nurs 1990; 9:39–45

231. Bolwerk CA: Effects of relaxing music onstate anxiety in myocardial infarction pa-tients. Crit Care Nurs Q 1990; 13:63–72

232. White JM: Effects of relaxing music on car-diac autonomic balance and anxiety afteracute myocardial infarction. Am J Crit Care1999; 8:220–230

233. Zimmerman L, Nieveen J, Barnason S, et al:The effects of music interventions on post-operative pain and sleep in coronary arterybypass graft (CABG) patients. Sch Inq NursPract 1996; 10:153–170

234. Labyak SE, Metzger BL: The effects of ef-fleurage backrub on the physiological com-ponents of relaxation: a meta-analysis. NursRes 1997; 46:59–62

235. Richards KC: Effect of a back massage andrelaxation intervention on sleep in criticallyill patients. Am J Crit Care 1998; 7:288–299

APPENDIX A—SUMMARY OFRECOMMENDATIONS

1. All critically ill patients should havethe right to adequate analgesia andmanagement of their pain. (Grade ofrecommendation � C)

2. Pain assessment and response totherapy should be performed regu-larly by using a scale appropriate tothe patient population and systemat-ically documented. (Grade of recom-mendation � C)

3. The level of pain reported by the pa-tient must be considered the currentstandard for assessment of pain and


response to analgesia whenever pos-sible. Use of the NRS is recom-mended to assess pain. (Grade of rec-ommendation � B)

4. Patients who cannot communicateshould be assessed through subjec-tive observation of pain-related be-haviors (movement, facial expres-sion, and posturing) andphysiological indicators (heart rate,blood pressure, and respiratoryrate) and the change in these pa-rameters following analgesic ther-apy. (Grade of recommendation �B)

5. A therapeutic plan and goal of anal-gesia should be established for eachpatient and communicated to allcaregivers to ensure consistent anal-gesic therapy. (Grade of recommen-dation � C)

6. If intravenous doses of an opioid an-algesic are required, fentanyl, hydro-morphone, and morphine are therecommended agents. (Grade of rec-ommendation � C)

7. Scheduled opioid doses or a contin-uous infusion is preferred over an “asneeded” regimen to ensure consis-tent analgesia. A PCA device may beutilized to deliver opioids if the pa-tient is able to understand and oper-ate the device. (Grade of recommen-dation � B)

8. Fentanyl is preferred for a rapid on-set of analgesia in acutely distressedpatients. (Grade of recommenda-tion � C)

9. Fentanyl or hydromorphone are pre-ferred for patients with hemodynamicinstability or renal insufficiency.(Grade of recommendation � C)

10. Morphine and hydromorphone arepreferred for intermittent therapybecause of their longer duration ofeffect. (Grade of recommenda-tion � C)

11. NSAIDs or acetaminophen may beused as adjuncts to opioids in se-lected patients. (Grade of recommen-dation � B)

12. Ketorolac therapy should be limitedto a maximum of five days, with closemonitoring for the development ofrenal insufficiency or gastrointesti-nal bleeding. Other NSAIDs may beused via the enteral route in appro-priate patients. (Grade of recommen-dation � B)

13. Sedation of agitated critically ill pa-tients should be started only after pro-viding adequate analgesia and treatingreversible physiological causes. (Gradeof recommendation � C)

14. A sedation goal or endpoint should beestablished and regularly redefined foreach patient. Regular assessment andresponse to therapy should be system-atically documented. (Grade of recom-mendation � C)

15. The use of a validated sedation as-sessment scale (SAS, MAAS, or VICS)is recommended. (Grade of recom-mendation � B)

16. Objective measures of sedation, suchas BIS, have not been completelyevaluated and are not yet proven use-ful in the ICU. (Grade of recommen-dation � C)

17. Midazolam or diazepam should beused for rapid sedation of acutely ag-itated patients. (Grade of recommen-dation � C)

18. Propofol is the preferred sedativewhen rapid awakening (e.g., for neu-rologic assessment or extubation) isimportant. (Grade of recommenda-tion � B)

19. Midazolam is recommended forshort-term use only, as it producesunpredictable awakening and time toextubation when infusions continuelonger than 48–72 hours. (Grade ofrecommendation � A)

20. Lorazepam is recommended for thesedation of most patients via inter-mittent i.v. administration or contin-uous infusion. (Grade of recommen-dation � B)

21. The titration of the sedative dose to adefined endpoint is recommended withsystematic tapering of the dose or dailyinterruption with retitration to mini-mize prolonged sedative effects. (Gradeof recommendation � A)

22. Triglyceride concentrations should bemonitored after two days of propofolinfusion, and total caloric intake fromlipids should be included in the nutri-tion support prescription. (Grade ofrecommendation � B)

23. The use of sedation guidelines, analgorithm, or a protocol is recom-mended. (Grade of recommenda-tion � B)

24. The potential for opioid, benzodiaz-epine, and propofol withdrawalshould be considered after highdoses or more than approximatelyseven days of continuous therapy.Doses should be tapered systemati-cally to prevent withdrawal symp-toms. (Grade of recommendation �B)

25. Routine assessment for the presence ofdelirium is recommended. (The CAM-ICU is a promising tool for the assess-ment of delirium in ICU patients.)(Grade of recommendation � B)

26. Haloperidol is the preferred agent forthe treatment of delirium in criti-cally ill patients. (Grade of recom-mendation � C)

27. Patients should be monitored forelectrocardiographic changes (QT in-terval prolongation and arrhythmias)when receiving haloperidol. (Gradeof recommendation � B)

28. Sleep promotion should include op-timization of the environment andnonpharmacologic methods to pro-mote relaxation with adjunctive useof hypnotics. (Grade of recommenda-tion � B)


Health & Medicine

Guia De Sedacion