INTRODUCTION — Distress generally presents as agitation. It is common among critically ill patients, especially those who are intubated or having difficulty communicating with their caregivers [1]. Distress needs to be treated for patient comfort and because it increases sympathetic tone, which may have untoward physiological effects [2]. Barring few exceptions (eg, neuromuscular paralysis, procedures) the administration of sedative-analgesic medications should not be based on anticipated distress but rather on that which is observed; otherwise, there will be an increased risk of over sedation which has been shown to worsen clinical outcomes.
The management of agitation in critically ill adults is reviewed here, including the initiation, maintenance, and withdrawal of pharmacological sedation. Common sedative-analgesic medications, the treatment of pain, and the use of neuromuscular blocking medications in critically ill patients are discussed elsewhere. (See "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects" and "Pain control in the critically ill adult patient" and "Neuromuscular blocking agents in critically ill patients: Use, agent selection, administration, and adverse effects".)
PRE-INITIATION — Before a sedative-analgesic agent is initiated to manage agitation, the cause of the distress should be identified and treated. Nonpharmacological strategies are preferred and should be implemented prior to the use of pharmacologic treatment.
Identify the cause of distress — Common causes of distress in critically ill patients include anxiety, pain, delirium, dyspnea, and neuromuscular paralysis. These etiologies may occur separately or in combination.
●Anxiety – Anxiety is defined as a sustained state of apprehension and autonomic arousal in response to real or perceived threats [1]. Fear of suffering, fear of death, loss of control, and frustration due to the inability to effectively communicate are typical causes of anxiety in critically ill patients. Symptoms and signs include headache, nausea, insomnia, anorexia, dyspnea, palpitations, dizziness, dry mouth, chest pain, diaphoresis, hyperventilation, pallor, tachycardia, tremulousness, and/or hypervigilance.
Identifying and treating the proximate cause of anxiety is always ideal as it may ameliorate both problems. Dyspnea, for example, is a common underlying cause of anxiety among critically-ill patients. Thus, if inadequate ventilator flow is causing dyspnea with resultant anxiety, the ultimate treatment for the anxiety (and underlying dyspnea) may be adjustment of ventilator settings. Alternatively, the abrupt onset of anxiety may prompt further work-up for a cardiopulmonary source.
●Pain – Routine patient care (eg, suctioning, repositioning, physical therapy), immobility, trauma, surgery, endotracheal tubes, and other monitoring devices can all produce pain. Evidence of pain may include grimacing, withdrawal, combativeness, diaphoresis, hyperventilation, and/or tachycardia. While self-report is preferred over behavioral pain scales, pain scales are superior to vital signs alone for assessment of pain [3]. The assessment and treatment of pain should be viewed as a priority in patients regardless of ability to communicate as pain is likely under reported during intensive care unit (ICU) stay. Upon ICU discharge, a significant proportion of patients report experiencing moderate to severe pain during ICU care [4]. (See "Pain control in the critically ill adult patient", section on 'Assessment for pain'.)
●Delirium – Delirium is an organic mental syndrome. It is defined as an acute and potentially reversible impairment of consciousness and cognitive function that fluctuates in severity [1]. Delirium occurs in up to 80 percent of ICU patients, but it is frequently unrecognized in older individuals and in patients with hypoactive delirium [5,6]. Delirium may be associated with an underlying cause such as infection, iatrogenic such as medications, or environmental. Prior to treatment, patients should be evaluated for a precipitating factor. In the acute phase, delirious patients have impaired short-term memory, abnormal perception, and intermittent disorientation, which is usually worse at night. Electroencephalography may display diffuse slowing. Delirium due to drug or alcohol withdrawal typically presents as a hyperactive delirium [7]. (See "Management of moderate and severe alcohol withdrawal syndromes", section on 'Delirium tremens' and "Diagnosis of delirium and confusional states", section on 'Evaluation'.)
Delirium is a risk factor for prolonged hospitalization and mortality in critically ill patients [8-10]. Risk factors for delirium include electrolyte imbalances (hypocalcemia, hyponatremia), hyperamylasemia, hyperglycemia, azotemia, hepatic disease (hyperbilirubinemia, elevated hepatic enzymes), infections, drug withdrawal, alcohol withdrawal, malnutrition, cancer, cerebrovascular disease, cardiopulmonary disease, advanced age, and some medications (benzodiazepines, corticosteroids, antihistamines, beta blockers, antiarrhythmics, digitalis glycosides, atropine) [11]. It has been suggested that "hypoactive delirium" be renamed "acute apathy syndrome" because it has a different constellation of causes than "hyperactive delirium" [12]. (See "Acute toxic-metabolic encephalopathy in adults".)
Society guidelines promote the regular assessment of critically ill patients with a delirium assessment tool such as the CAM-ICU or the Intensive Care Delirium Screening Checklist (ICDSC) [13], understanding that the level of arousal may affect the assessment [3], which is discussed separately. (See "Diagnosis of delirium and confusional states", section on 'Recognizing the disorder'.)
●Dyspnea – Dyspnea is a sensation of air hunger or a feeling of suffocation [1]. Evidence of dyspnea may include tachypnea, shallow breathing, diaphoresis, tachycardia, use of the accessory muscles of respiration, hypoxemia, and/or hypercapnia. Dyspnea may exist despite acceptable blood gas parameters. Strategies to alleviate hypoxia or dyspnea such as adjustment of ventilator settings, if possible, should be explored prior to the use of medications.
●Neuromuscular paralysis – All patients undergoing neuromuscular blockade require pharmacological sedation, since neuromuscular paralysis without sedation or adequate pain control is an extremely frightening and unpleasant sensation. Identifying distress in patients undergoing neuromuscular blockade is difficult because the typical physiological responses associated with stress (eg, increased heart rate and blood pressure with stimulation) may not correlate with patient discomfort in this setting. (See "Neuromuscular blocking agents in critically ill patients: Use, agent selection, administration, and adverse effects".)
Treat the cause of distress — Initial treatment of agitation should target the presumed cause of the underlying distress. As an example, a patient who is agitated due to hypoxemia should receive supplemental oxygen.
Nonpharmacological strategies — Nonpharmacological strategies for managing agitation should begin simultaneously with therapy targeting the cause of distress (since the cause of the distress is rarely quickly reversible). Nonpharmacological strategies include reassurance, frequent communication with the patient, regular caregiver visits, establishment of normal sleep cycles, and cognitive-behavioral therapies [14]. Examples of cognitive-behavioral therapies include music therapy, guided imagery, and relaxation therapy.
This strategy of using nonpharmacological interventions to prevent or control agitation, rather than immediately initiating pharmacological sedation, is supported by evidence:
●One trial randomly assigned 140 mechanically ventilated patients to receive either a strategy of no sedation followed by continuous verbal comforting and reassurance or continuous sedation with daily interruption [15]. Only when nonpharmacologic interventions failed were patients treated with a continuous sedative infusion with daily interruption. The trial found that patients managed with a strategy of no sedation had more ventilator-free days and a decreased length of ICU stay, length of hospital stay, and incidence of delirium. There was no difference in posttraumatic stress disorder, quality of life, depression, or recall of the ICU experience in survivors approximately two years after randomization [16].
●Another randomized multicenter trial of 373 mechanically ventilated patients reported that, compared to usual care (UC) or noise-cancelling headphones (NCH), patient directed self-initiated music (PDM; with preferred selections tailored by a music therapist) was associated with a reduction in the visual analog scale for anxiety (52 versus 33) throughout the study period (up to 5.7 days) [17]. In addition, by the fifth day, PDM was associated with a reduction in sedation intensity (sedation intensity score 4.4 versus 2.8) and frequency (5 doses versus 3 doses of study-approved sedative). These findings were confirmed in a meta-analysis of 14 trials examining the impact of music in mechanically ventilated patients [18] and reductions in patients' distress was also reported in ICU patients in the ICU who were receiving noninvasive ventilation [19].
While physical restraints are used by some ICUs, they should never be the sole method employed for managing patients in the ICU. Their use should be supplementary to other more acceptable methods of sedation and use should be transient. When employed, efforts should be made to reduce the use of physical restraints in the ICU.
INITIATION — Sedative-analgesic medication is indicated when treatment of the cause of the distress and nonpharmacological interventions cannot sufficiently control the agitation. The Society of Critical Care Medicine has published guidelines regarding the selection and initiation of sedative-analgesic medications in critically ill patients [3].
Available agents — Sedative-analgesic medications that are commonly used in the intensive care unit (ICU) include benzodiazepines (eg, diazepam, lorazepam, midazolam), opioid analgesics (eg, fentanyl, hydromorphone, morphine, remifentanil), propofol, dexmedetomidine, ketamine, and antipsychotics (eg, haloperidol, quetiapine, ziprasidone) (table 1) [7,20]. Agents such as acetaminophen, nonsteroidal antiinflammatory drugs (ketorolac), gabapentin or pregabalin, and antiepileptics can be used as adjunctive therapy when indicated. All these agents differ in their amount of anxiolysis, analgesia, amnesia, and hypnosis (table 2). Their mechanisms, properties, dosage regimens, and potential adverse effects are reviewed separately (table 3). (See "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects".)
Although barbiturates (eg, thiopental, methohexital) can be used to manage agitation during critical illness in patients not tolerating or responding to other agents, it is not ideal. This is because barbiturates are not potent sedatives and can cause profound cardiovascular and respiratory depression, as well as diminished cerebral blood flow. Phenobarbital may be useful in patients with alcohol withdrawal. Sevoflurane (a polyfluorinated methyl-isopropyl volatile anesthetic) is a newer agent for which safety and efficacy data are limited, prohibiting its routine use as a sedative in the ICU [21-23]. Fospropofol is not available worldwide and thiopental is no longer manufactured in United States or Canada.
Selection of an agent — No sedative-analgesic agent is sufficiently superior to other agents to warrant its use in all clinical situations. The Society of Critical Care Medicine guidelines favor nonbenzodiazepine agents due to evidence of shorter duration of mechanical ventilation, but the optimal agent for short-term or long-term therapy is not known [3]. Selection of an agent must be individualized according to patient characteristics and the clinical situation [24]. Important considerations when selecting a sedative-analgesic agent include the etiology of the distress, expected duration of therapy, clinical status of the patient, and potential interactions with other drugs:
●Etiology of the distress – The appropriate initial pharmacological agent for managing agitation due to distress depends upon the cause of the distress [3]:
•For distress due to dyspnea or pain, opioids are the agents of choice. (See "Pain control in the critically ill adult patient".)
•For distress due to delirium, antipsychotics (eg, haloperidol, quetiapine, ziprasidone) should not be routinely used but may be used in cases of significant distress induced by delirium. Although weak evidence suggests that dexmedetomidine may decrease the risk of delirium [25], an Intensive Care Medicine Rapid Practice Guideline (ICM-RPG) has suggested that dexmedetomidine be used when clinicians feel that a reduction in delirium risk outweighs the risk of adverse effects of hypotension and bradycardia [26]. To date there are no agents that prevent delirium. Therapies that do not work include statins and bright light. (See "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects", section on 'Antipsychotics'.)
•For agitation due to stress or anxiety, the Society of Critical Care Medicine endorses the use of propofol rather than benzodiazepines in cardiac surgery patients and propofol or dexmedetomidine rather than benzodiazepines in other surgical and medical patients [3]. (See "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects", section on 'Propofol' and "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects", section on 'Dexmedetomidine' and "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects", section on 'Benzodiazepines'.)
However, combination therapy is common in the ICU since many patients have more than one cause of distress. As an example, a benzodiazepine plus an opioid is appropriate for a patient whose agitation is due to anxiety and pain. For patients who are intubated and mechanically ventilated and not able to clearly communicate the source of agitation, analgesia should always be provided first [3]. (See "Pain control in the critically ill adult patient".)
●Pharmacokinetic modifying variables (eg, age, body weight, renal and hepatic function) and the desired depth of sedation should also be considered whenever a sedative-analgesic agent is selected. Abnormal pharmacokinetic modifying variables can magnify differences among the sedative-analgesic agents (eg, onset, peak, duration of sedation), especially during deep or long-term sedation.
Initial dose — The initial dose of a sedative-analgesic agent should account for the desired level of sedation ability to tolerate the drug including hemodynamic and respiratory status as well as factors that may affect drug metabolism (ie, patient age, body weight, renal function, hepatic function, drug interactions, history of alcoholism, history of drug abuse). Higher doses are appropriate for deeper sedation and larger patients, while lower doses are appropriate for lighter sedation, smaller patients, patients with advanced age, diminished renal function, or decreased hepatic function. Patients with a history of alcohol or abuse may require higher doses of benzodiazepines or opioids, respectively, to achieve a given effect. (See "Sedative-analgesic medications in critically ill adults: Properties, dose regimens, and adverse effects".)
Administration — The evidence indicates that continuous infusion of a sedative-analgesic prolongs the duration of mechanical ventilation [27]. As a result, current practice favors intermittent bolus doses, daily interruption or dose minimization titrated to light level of sedation (RASS -2 to 0) of continuous infusions [3]. Clinical practice guidelines for the sustained use of sedatives and analgesics in critically ill adults endorse the initial use of intermittent bolus dose, with the initiation of continuous infusions with daily interruption or dose minimization titrated to light level of sedation in patients who require intermittent infusions more often than every two hours [7].
Sedation goal — The ideal sedation goal is for the patient to be awake and comfortable with minimal to no distress (eg, 0 on the RASS scale) (table 4), although some patients may require a deeper level of sedation for optimal management. In general, light levels of sedation are promoted because they decrease mechanical ventilation days and tracheostomy rates, although no effect on mortality has been demonstrated [3]. The sedation goal should be ascertained at the bedside for each patient; while the Society of Critical Care Medicine (SCCM) promotes light sedation rather than heavy sedation, they recognize that no universally accepted definition of light sedation exists. We also agree with the 2018 SCCM [3] guidelines which stated that targeting a RASS level of -2 (previously promoted by 2013 guidelines [7]) is too deep and could lead to the over sedation and delayed recovery of many patients (eg, by eliminating the opportunity for daily physical therapy). In contrast, a proportion of critically-ill, mechanically-ventilated patients require a very deep level of sedation-analgesia to control agitation or pain. A patient-centered approach, derived by the bedside clinician, is best and should be used to determine the appropriate goal for the depth of sedation. This goal should be determined prior to beginning or escalating sedative-analgesic medications, since this is the target to which initial therapy is titrated. As examples, lighter sedation may be desired when serial neurological exams are required, while deeper sedation may be desired during severe hypoxemic respiratory failure.
The goal depth of sedation should be frequently reassessed and adjusted as the patient's sedation requirement becomes more apparent. Some patients require no sedation, while others require deep sedation to be mechanically ventilated without discomfort, agitation, or asynchrony.
The depth of sedation is typically assessed during the initiation of pharmacological sedation using criteria such as spontaneous movement and response to verbal and tactile stimuli. Scoring systems have been developed to make evaluation of the depth of sedation more rigorous and quantitative, but these systems are more commonly used during the maintenance of sedative-analgesic therapy, as described below. (See 'Monitoring' below.)
MAINTENANCE — Once the initiation of the sedative-analgesic agent achieves a calm state, attention should be directed toward monitoring and avoiding excess sedation. This involves frequent reassessment of pain and sedative requirements to achieve simultaneous patient comfort in an awake and alert patient.
Monitoring — The maintenance of pharmacological sedation requires that patients be reassessed frequently to determine whether their agitation and underlying distress are being adequately managed. Scoring systems have been developed to facilitate this evaluation.
There are scoring systems (ie, scales) to assess pain, sedation, and delirium. The scale appropriate for the presumed cause of distress should be employed. As an example, if the distress was felt to be due to pain and an opioid was initiated, then assessment using a pain scale is appropriate. If the goal of therapy is sedation, then a scale assessing level of sedation should be used.
Once the appropriate scoring system has been used to determine whether the agitation and/or underlying distress is sufficiently controlled, the sedative-analgesic medication should be titrated or tapered to meet the therapeutic goals.
Scoring systems — Scoring systems use multiple criteria to determine the amount of pain, depth of sedation, or presence or severity of delirium. An important limitation of the scoring systems is that reference standards do not exist [28]:
●Pain scales – There exist unidimensional scales (ie, verbal rating scale, visual analogue scale, numeric rating scale) and multidimensional scales (ie, McGill Pain Questionnaire, Wisconsin Brief Pain Questionnaire) to assess a patient's level of pain. The unidimensional scales can be quickly and easily applied in the intensive care unit if the patient is communicative. As an example, the numeric rating scale is a zero to ten point scale on which ten represents the worst pain. Patients choose the number that best describes their pain. The multidimensional scales are more complex and take longer to administer; thus, they may not be appropriate for the intensive care unit. The Behavioral Pain Scale and the Critical-Care Pain Observation Tool are simple, valid, and reliable pain scales for pain assessment in critically ill patients and are recommended as scales of choice by current guidelines [7].
●Sedation scales – There are numerous scoring systems to assess the depth of sedation that are valid and reliable in adults who are mechanically ventilated and critically ill [29-33]. Current guidelines support the use of the Richmond Agitation-Sedation Scale (RASS) (table 4) and the Riker Sedation-Agitation Scale (SAS) [7]. Alternative scoring systems include the Motor Activity Assessment Scale (MAAS), Minnesota Sedation Assessment Tool (MSAT), Ramsay Sedation Scale (table 5), Bizek Agitation Scale, Sheffield Scale, and COMFORT Scale [34-37]. The COMFORT scale is a valid and reliable system for children [35].
●Delirium scales – Many scales and diagnostic instruments have been developed to identify and evaluate delirium, but most exclude critically ill patients due to difficulty communicating with them [3]. However, a rapid bedside instrument that can identify delirium in critically ill patients is the Confusion Assessment Method for the ICU (CAM-ICU) [38]. The Intensive Care Delirium Screening Checklist (ICDSC) is also a simple and valid tool for bedside assessment of delirium. Both scales assess patients for acute mental status changes or fluctuating mental status changes, inattention, disorganized thinking, and/or an altered level of consciousness. Both the CAM-ICU and the ICDSC can identify new or persistent delirium, but neither quantify the severity of the delirium. (See "Diagnosis of delirium and confusional states", section on 'Recognizing the disorder'.)
Bispectral index — For patients who are pharmacologically paralyzed, monitoring is challenging because the scoring systems cannot determine the level of pain, depth of sedation, or presence of delirium. Heart rate and blood pressure have historically been used as indicators of distress in this situation, but these vital signs are neither sensitive nor specific. We believe that there are two reasonable approaches. Pharmacologically paralyzed patients can be given higher than usual doses of both an anxiolytic/amnestic and an analgesic to ensure deep sedation. Alternatively, bispectral index (BIS), auditory evoked potentials, or other objective monitoring systems can be used (table 6) [3]. The latter approach may limit drug accumulation.
BIS monitoring uses Fourier transform analysis of electroencephalographic data to estimate the depth of sedation. It is used primarily during operative anesthesia in patients without underlying neurologic disease. BIS monitoring is not used routinely in the ICU because there are conflicting data regarding its benefit and electromyelographic activity from the scalp muscles creates artifact [39-45]. We and others believe BIS monitoring is a reasonable approach to assessing depth of sedation in ICU patients receiving deep levels of sedation or neuromuscular paralysis [3]; however, we agree that BIS monitoring should not replace the clinical assessment of sedation in the routine management of ICU patients until more favorable data are reported. Also, clinicians should recognize the potential for misinterpretation of the BIS algorithm in patients with electromyographic activity [46].
Avoid excess sedation — Sedative-analgesic medications should not be overused because excess sedation may unnecessarily prolong the duration of mechanical ventilation [27,47,48]. Two strategies have been shown in randomized trials to reduce duration of mechanical ventilation and complications related to prolonged mechanical ventilation: intermittent boluses of medication (including analgesia without sedatives, sometimes referred to as "no" sedation) [27,49] and the daily interruption of continuous infusions [50-53]. Both of these approaches have been protocolized in many ICUs in an attempt to avoid excess sedation; however, the value of protocols in this regard remains unproven [54,55]. We believe there is sufficient evidence to justify efforts to minimize sedative-analgesic infusions, although the optimal method (eg, protocol-driven intermittent infusions, daily interruption, or a combination) is not known. Additional studies focusing on efficacy, feasibility, and safety are needed to determine the optimal approach.
Intermittent bolus — An observational study of 242 patients compared the duration of mechanical ventilation among patients who received a continuous sedative-analgesic infusion to those who received either intermittent sedative-analgesic infusions or no sedative-analgesics based on a nursing protocol [49]. The group that received intermittent infusions or no medication had a shorter duration of mechanical ventilation (median of 56 hours) than the group that received a continuous infusion (median of 185 hours). Data that compared no sedation to light sedation with daily interruption of sedative infusions are discussed below. (See 'Daily interruption and nursing protocolized sedation' below.)
Daily interruption and nursing protocolized sedation — Daily interruption of sedation (DSI) refers to discontinuing the continuous sedative-analgesic infusion until the patient is awake and following instructions, or until the patient is uncomfortable or agitated, and deemed to require the resumption of sedation. The rationale DSI is that they facilitate assessment of the patient's underlying neurologic status, as well as the patient's need for ongoing sedation. Nursing-protocolized (NP)-sedation is defined as an established sedation protocol implemented by nurses at the bedside to sedative choices and medication titration to achieve prescription-targeted sedations scores. Several studies have compared these mechanisms and no consistent robust difference has been reported [51,56-58]. The SCCM state that light sedation can be achieved in most patients most of the time using either method. Thus, many ICUs practice one or both of these methods.
Randomized trials and meta-analyses report possible benefit from DSI or NP-sedation with regard to reducing duration of mechanical ventilation and length of stay. However, there is considerable heterogeneity among trials which limits interpretation of the analysis. As examples:
●In a trial of 128 patients who were receiving mechanical ventilation and a continuous sedative-analgesic infusion, patients were randomly assigned to continue conventional management or to undergo daily spontaneous awakening trials [50]. The spontaneous awakening trials consisted of interruption of the continuous infusion until the patient was awake. The group whose continuous infusion was interrupted daily had a shorter duration of mechanical ventilation (4.9 versus 7.3 days) and length of ICU stay (6.4 versus 9.9 days), as well as fewer neurodiagnostic tests. Limitations of the trial include that it was performed in a single center, ventilator weaning was not standardized, and the spontaneous awakening trials were monitored closely by study personnel, which is not feasible in most ICUs.
●A similar trial randomly assigned 336 patients to a daily spontaneous breathing trial and either a daily spontaneous awakening trial or conventional sedation management [59]. The daily spontaneous awakening trial group had decreased one-year mortality (but not 28-day mortality), an increased number of ventilator-free days, a decreased length of ICU stay, and a decreased length of hospital stay. The group also had less cognitive impairment at three months (absolute risk reduction of 20 percent), although there was no difference at 12 months [60].
●A meta-analysis of nine trials demonstrated only marginal reductions in the duration of mechanical ventilation (13 percent), ICU and hospital length of stay (10 and 6 percent, respectively) when compared with strategies that do not utilize daily sedative interruption [61]. There was no difference in risk of death, rate of accidental endotracheal tube removal, incidence of new onset delirium, or in the doses of sedative administered. The confidence intervals were wide which indicates imprecision and limits interpretation of the analysis.
●A meta-analysis of six randomized trials of mechanically ventilated patients in closed nonspecialty ICUs reported that, compared with usual care, protocolized sedation (algorithm and/or daily interruption) was associated with a rate reduction in overall mortality (15 percent), length of hospital stay (3.5 days), and tracheostomy (31 percent) [62]. There was no difference on the duration of mechanical ventilation and rate of self-extubation or re-intubation.
While these trials indicate that daily interruption of continuous sedative-analgesic infusions are beneficial, compliance to protocols can be challenging, and the benefit may be less or absent if a sedation protocol is also being concurrently used [63-65]. This was illustrated by a multicenter trial in which 430 mechanically-ventilated patients were randomly assigned to receive protocolized sedation (PS) alone or protocolized sedation plus daily interruption (PS+DI) of their continuous sedative-analgesic infusion [56]. The median time to successful extubation, the primary outcome, was seven days in both groups. The PS+DI group had higher mean doses of infusions of midazolam and fentanyl as well as a greater number of boluses of benzodiazepines and opiates, than the PS group. No differences were noted in the rates of unintentional removal of medical devices, ICU delirium, diagnostic neuroimaging, or tracheostomy. Of note, the average daily dose of midazolam was higher in both groups in this study than in the daily interruption trial described above [50]. The seemingly disparate results from the two clinical trials may be reconciled by the notion that protocolized weaning of sedation by the bedside nurse (as in the control group of the PS versus PS+DI study [56]) can effectively achieve the minimal effective sedative dose requirement for patients. This would explain the lack of any further benefit, when the minimal effective dose is interrupted.
Daily interruption of light sedation has been compared with "no sedation." In a single-center randomized trial of 710 mechanically ventilated patients, light sedation, defined as a score of -2 to -3 on the Richmond Agitation and Sedation Scale (RASS; this scale ranges from -5 [unresponsive] to +4 [combative]) with daily interruption, was compared with "no sedation" [66]. Patients in the sedation group received a propofol infusion followed by midazolam while patients in the "no sedation" group were allowed to receive intermittent analgesia with intravenous morphine and sedated only if necessary and if nonpharmacologic methods and analgesia with morphine had failed. By day 7, the mean RASS score was -0.8 in the nonsedation group and -1.8 in the sedation group. Although the 90-day mortality was lower in the sedation group, it was not statistically significant (37 versus 42 percent). The number of ICU- and ventilator-free days and days free from coma or delirium was also no different. Although the thromboembolic event rate was higher in the sedation group, the total event rate was low (11 of 710 patients [1.5] percent). However, the inclusion criteria were very specific (among 2300 screened only 710 underwent randomization) suggesting that the approach of "no sedation" is not generalizable to all mechanically ventilated patients. In addition, one-third of patients in the nonsedation group received sedation in the first 24 hours after randomization and continued to receive intermittent morphine during the trial. Furthermore, the trial may have been underpowered to detect a difference in mortality. Further studies are required before the practice of "no sedation" can be routinely applied to mechanically ventilated patients in the ICU.
Concerns related to patient safety have been a significant obstacle to implementation of daily interruption of continuous sedative-analgesic infusions [67,68]. These concerns include the possibility of long-term psychological sequelae (eg, posttraumatic stress disorder [PTSD]) and myocardial ischemia. There is little evidence to support these concerns, as demonstrated by the following studies:
●An observational study performed using patients from the first randomized trial described above plus contemporaneous patients that were not enrolled in that trial found that patients who received daily interruption of their continuous sedative-analgesic infusion did not experience adverse psychological outcomes and were less likely to have symptoms of PTSD than those who received conventional management [69].
●In the trial described above that randomly assigned 336 patients to a daily spontaneous breathing trial and either a daily spontaneous awakening trial or conventional sedation management [59], there was no difference in the frequency of PTSD at 3 or 12 months [60].
●A prospective cohort study evaluated 74 patients with risk factors for coronary artery disease who were receiving mechanical ventilation [70]. Electrocardiographic monitoring was performed during the continuous sedative-analgesic infusion and during interruption of the continuous infusion. Myocardial ischemia (defined as ST segment elevation or depression >0.1 mV from baseline lasting 10 minutes or longer) was identified in 24 percent of the patients at some time during the study. Myocardial ischemia was not more common during interruption of the continuous infusion, although heart rate, blood pressure, respiratory rate, and catecholamines increased significantly. The study did not address the cumulative effects of multiple days of interrupting the continuous infusion.
WITHDRAWAL — When pharmacological sedation is no longer necessary, the sequence and rate of discontinuing the sedative-analgesic agents must be determined:
●For patients receiving more than one sedative-analgesic medication (eg, a sedative and an opioid), the opioid should be tapered last so that the patient does not awake in pain.
●The rate of the reduction should be individualized. Generally speaking, discontinuation over a short period of time (eg, hours) is acceptable if the sedative-analgesic agent has been administered for a short duration (≤7 days). In addition, abrupt discontinuation may be appropriate in patients who have received sedation for greater than seven days who are deeply sedated from prolonged accumulation of medication. However, a gradual reduction (~10 to 25 percent per day) may be necessary if the sedative-analgesic agent has been administered for >7 days and the patient exhibits evidence of tachyphylaxis, with increasing dosage required over time to achieve the same level of sedation.
It is important for clinicians to realize that there may be a delay (ie, days) between the moment that reduction of the sedative-analgesic agent begins and the patient begins to awaken, particularly following long-term therapy. This is because lipophilic drugs accumulate in tissue stores and must be mobilized for elimination.
During the reduction of the sedative-analgesic medication, the patient should be closely observed for withdrawal symptoms. Acute withdrawal symptoms in this setting appear to be common. In an observational study of 28 mechanically ventilated patients who had been in the ICU for greater than one week, nine patients (32 percent) developed acute withdrawal symptoms when their sedative-analgesic medication was reduced [71]. Higher doses of benzodiazepines and opiates conferred a higher risk of withdrawal.
Benzodiazepine withdrawal symptoms include agitation, confusion, anxiety, tremors, tachycardia, hypertension, and fever. Seizures may also occur. The administration of intermittent intravenous or oral lorazepam (0.5 to 1 mg every 6 to 12 hours) may help protect the patient from developing withdrawal symptoms as the continuous benzodiazepine infusion reduced.
Opiate withdrawal symptoms include agitation, anxiety, confusion, rhinorrhea, lacrimation, diaphoresis, mydriasis, piloerection, stomach cramps, diarrhea, tremor, nausea, vomiting, chills, tachycardia, hypertension, and fever. Several strategies have been proposed for preventing opioid withdrawal, including de-escalating the dose, converting to a longer acting oral equivalent, converting to a long-acting barbiturate (eg, phenobarbital), and adding an alpha-2-agonist (clonidine, dexmedetomidine) [72,73]. However, there are no controlled trials of any strategy and there is no consensus as to the best strategy. Data are limited to case reports, including two reports in which dexmedetomidine was initiated at a dose of 0.7 mcg/kg/hour (with or without a loading dose) and successfully facilitated opioid withdrawal [74,75].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Nonprocedural sedation".)
SUMMARY AND RECOMMENDATIONS
●Rationale – Distress is common in critically ill patients, particularly among those who are intubated or have difficulty communicating with their caregivers. However, many patients may be comfortable without requiring a depressed level of consciousness. (See 'Introduction' above.)
●Assessment – Before a sedative-analgesic agent is initiated to treat agitation due to distress, the cause of the distress should be identified and treated. Nonpharmacological strategies should be implemented simultaneously. (See 'Pre-initiation' above.)
●Initiation – Pharmacological sedation is indicated when treatment of the cause of the distress and the nonpharmacological interventions cannot sufficiently control the agitation. (See 'Initiation' above.)
•Agent selection – Available agents are listed in the table (table 1). The sedative-analgesic agent and its initial dose are selected on the basis of several factors: the etiology of the distress, the expected duration of therapy, potential drug interactions, the desired depth of sedation, and pharmacokinetic-modifying variables (table 3). (See 'Available agents' above and 'Selection of an agent' above and 'Initial dose' above.)
•Administration – We recommend not routinely using uninterrupted sedative-analgesic infusions to sedate critically ill patients (Grade 1B). Intermittent boluses, continuous infusions directed by a sedation protocol, or daily interruption of continuous infusions are preferable. (See 'Administration' above and 'Avoid excess sedation' above.)
●Monitoring – All patients should be frequently reassessed to determine whether their agitation and underlying distress are being adequately managed (table 4 and table 5). The sedative-analgesic medication should then be titrated or tapered accordingly. Scoring systems have been developed to facilitate the evaluation. (See 'Monitoring' above.)
●Withdrawal – Once comfort is achieved initially, the goal depth of sedation should be frequently reassessed and adjusted. As the respiratory failure or other critical illness is treated, the sedative requirement should fall. Thus, daily attempts should be made to reduce the level of sedation, although the rate of the reduction must be individualized. For patients receiving more than one sedative-analgesic medication, we taper the opioid last, so that the patient does not awake in pain. During the reduction of sedation, the patient should be closely observed for withdrawal symptoms. (See 'Withdrawal' above.)
