INTRODUCTION — Distress due to pain, fear/anxiety, dyspnea, or delirium is common among critically ill patients, especially those who are intubated or are having difficulty communicating with their caregivers [1]. Distress may manifest clinically as agitation that is often associated with ventilator asynchrony and vital sign abnormalities. Regardless, distress needs to be treated to comfort the patient, ameliorate agitation that may interfere with supportive care, and attenuate increases in sympathetic tone, which may have untoward physiologic effects [2].
Common sedative-analgesic medications used to treat distress in critically ill adults are reviewed here. Identifying the cause of distress and using this information to select the optimal sedative-analgesic agent are discussed separately. (See "Sedative-analgesic medications in critically ill adults: Selection, initiation, maintenance, and withdrawal".)
The management of pain and neuromuscular blockade in critically ill patients are also described separately. (See "Pain control in the critically ill adult patient" and "Neuromuscular blocking agents in critically ill patients: Use, agent selection, administration, and adverse effects".)
ANALGESICS — Pain can be managed in the intensive care unit (ICU) with opioid analgesic and nonopioid analgesic agents. The choice of agent should be directed at the etiology of pain, but opioids are, in general, frequently administered for the management of pain in mechanically ventilated patients.
Opioid analgesics — Morphine sulfate, fentanyl, and hydromorphone are the intravenous opioids that are most commonly used to manage distress due to pain in critically ill patients. Oral opioids such as oxycodone, methadone, and morphine are also given to patients where oral or enteral administration is preferred. Remifentanil is also an option; advantages include its rapid onset of action and prompt clearance that are independent of hepatic and renal function, although there is debate as to whether its use is associated with a higher risk of tolerance [3-5].
All of the opioids lack amnestic properties, induce tolerance, and have similar analgesic and sedative properties when administered in equipotent doses [6]. Specific agents differ in their onset, duration of action, histamine-related side effects, and risk of accumulation in organ failure. The pharmacologic properties of the opioids and principles for selection of intermittent, preprocedural, continuous, and patient-controlled dosing are described in detail elsewhere. (See "Pain control in the critically ill adult patient", section on 'Opioid analgesics'.)
The Society of Critical Care Medicine 2018 guidelines place emphasis on the administration of intravenous opioids for non-neuropathic pain and analgesia-first sedation for the management of agitation in mechanically ventilated ICU patients [7]. The comparative advantages and disadvantages and typical dose regimens, as well as the roles of analgesics, sedatives, anxiolytics, and antipsychotics in critically ill patients, are described in the table (table 1). Limited information is available regarding drug dosing in patients with obesity who are critically ill [8]. Dosing for patients with obesity is described separately. (See "Intensive care unit management of patients with obesity".)
Numerous drugs used commonly in the ICU have the potential to interact with opioids. Central nervous system (CNS) and respiratory depressants (eg, benzodiazepines) enhance the CNS and respiratory depressant effect of opioids, while CNS stimulants (eg, methylphenidate) decrease the CNS depressant effect of opioids. Other drug interactions are related to the metabolism of opioids via CYP3A4. As examples, the azole antifungals (eg, fluconazole, itraconazole, posaconazole, ketoconazole, voriconazole) and the macrolides and related antibiotics (eg, clarithromycin, erythromycin, telithromycin) may prolong fentanyl activity by inhibiting CYP3A4. The rifamycins (eg, rifampin, rifabutin) may decrease the serum concentration and effects of opioids.
The ultrashort-acting opioids alfentanil and sufentanil are generally not used in critically ill patients, because they do not offer advantages and may be more costly. The long-acting opioid agonist-antagonist analgesics are also not used very often, because they are long-acting and may not produce sufficient analgesia due to ceiling effects. This class of agents includes buprenorphine, butorphanol, dezocine, nalbuphine, and pentazocine.
Nonopioid analgesics — Acetaminophen (enteral or intravenous), nonsteroidal antiinflammatory agents, and intravenous ketamine can be used as alternatives or additives to opioids for the management of nonneuropathic pain in the ICU (table 1). Additionally, they can be used as opioid-sparing agents to reduce or eliminate the need for opioids in adult ICU patients [7]. Enteral gabapentin, pregabalin, and/or carbamazepine, in addition to intravenous opioids, can be used to treat neuropathic pain in adult ICU patients. (See "Pain control in the critically ill adult patient", section on 'Nonopioid analgesics'.)
PROPOFOL — Propofol is an intravenous anesthetic that is commonly used for sedation of the agitated adult intensive care unit (ICU) patient. It is particularly useful when rapid sedation and rapid awakening is desirable (eg, patients who require frequent neurologic examinations) because it has a short duration of effect. In one large, randomized, open label trial, compared with intermittent bolus lorazepam (median dose 11.5 mg/day), propofol infusions with daily interruption (mean dose 24.4±16.3 mcg/kg/min) resulted in significantly lower number of mechanical ventilation days but did require higher morphine doses per ventilator day [9]. A multicenter ICU database analysis of over 3000 patients reported that, when compared with midazolam or lorazepam, propofol infusions were associated with a lower mortality (risk ratio 0.76 and 0.78, respectively), earlier discharge, and discontinuation from mechanical ventilation [10]. Data from comparative trials with current sedation guidelines are needed to confirm these results.
Mechanism — Activation of the central gamma-aminobutyric acid receptors (GABA[A] receptors) with modulation of hypothalamic sleep pathways appear to be the mechanism by which propofol exerts its effect [11-14].
Properties — Propofol is a highly lipophilic phenol derivative that is insoluble in water. Therefore, it is formulated as an emulsion of soybean oil, egg lecithin, and glycerol for intravenous administration. Although labeled contraindications include hypersensitivity to eggs, egg products, soy, or soy products, reviews suggest a need for further evaluation of this issue [15,16]. Propofol has amnesic, anxiolytic, anticonvulsant, and muscle relaxant (including bronchodilation) effects. It has no direct analgesic effects.
●Onset of action – Propofol has an onset of action of less than one minute because its high lipophilicity facilitates passage through the blood-brain barrier.
●Duration of effect – Propofol's duration of effect is 3 to 10 minutes during short-term use (<48 hours). The short duration reflects the rapid metabolism of propofol by the liver and elsewhere to minimally active metabolites, which are renally excreted. Less is known about propofol's duration of effect following long-term administration. One study reported an elimination half-life of 31 hours after prolonged administration, suggesting that propofol's high lipophilicity leads to its accumulation in fatty tissues and prolonged sedation [17]. These data, however, may overestimate the magnitude of the sedative prolongation. Another study found that, when compared with patients receiving short-term sedation with propofol (<24 hours), the mean recovery time to extubation in those that received propofol for extended periods categorized as medium (1 to 7 days, n = 16) and long term (>7 days; n = 10) was delayed by only 10 and 30 minutes, respectively [18].
The elimination of propofol is not impaired by hepatic or renal dysfunction. Propofol has a large volume of distribution and is highly protein bound.
Dose regimens — A typical dose regimen for propofol is listed in the table (table 1) [7]. A loading dose is not typically administered when a propofol infusion is started in the ICU for sedation unless a bolus dose is required for emergency care.
The US Food and Drug Administration emergency use authorization to temporarily permit use of a product containing propofol 2% (propofol 20 mg/mL) emulsion during the coronavirus disease 2019 (COVID-19) pandemic was revoked (July 2022), and the product is no longer approved for use.
Administration — Propofol is administered by continuous infusion in the ICU and not by intermittent infusion, because it is associated with dose- and rate-dependent hypotension. This was best illustrated by an observational study of 25,981 patients receiving propofol anesthesia: 4079 patients (15.7 percent) developed hypotension, defined as a systolic blood pressure <90 mmHg [19]. Among those who became hypotensive, 77 percent of the hypotensive episodes occurred within 10 minutes of induction via a bolus infusion.
The manufacturer recommends that the bottles and tubing be discarded every twelve hours and that line integrity be maintained to minimize the risk of bacterial contamination.
When administered peripherally, propofol is generally given through a large-bore intravenous catheter (often in the antecubital fossa) to reduce burning, stinging, and pain that can occur with peripheral administration.
During administration, routine biochemical monitoring (triglycerides, serum lactate, creatinine kinase, myoglobin) may allow early identification of the propofol infusion syndrome (PRIS) [20]. Propofol should be discontinued as soon as any abnormality is identified. (See 'Propofol-related infusion syndrome' below.)
Adverse effects — Hypotension is a common consequence of propofol infusion, estimated to occur in approximately 25 percent of ICU patients who receive propofol for sedation [21]. Other potential adverse effects of propofol which are uncommon include bradycardia, arrhythmias, neuroexcitatory effects (seizure-like activity, myoclonus, choreoathetoid movements, meningismus), infections from contaminated vials or tubing, respiratory acidosis, pancreatitis, hypertriglyceridemia, anaphylaxis, and green or white discoloration of urine [17,22-25].
Unusual and potentially serious complications are associated with continuous infusion of propofol for longer than 24 to 48 hours. These include progressive hypertriglyceridemia, pancreatitis, increased carbon dioxide production, and an excessive caloric load (the emulsion contains approximately 1.1 kcal/mL, most of which is derived from lipids). Despite these complications, continuous infusion of propofol for greater than 24 hours does not appear to increase overall mortality, according to a meta-analysis of 14 randomized trials (1184 patients) [26].
Propofol-related infusion syndrome — PRIS is a rare complication of propofol infusion. It is generally associated with high doses (>4 mg/kg/hour or >67 mcg/kg/min) and prolonged use (>48 hours) [27-30], though it has been reported with high-dose short-term infusions [31,32]. Additional proposed risk factors include a young age, critical illness, high fat and low carbohydrate intake, inborn errors of mitochondrial fatty acid oxidation, and concomitant catecholamine infusion or steroid therapy [32]. Characteristics of PRIS include acute refractory bradycardia, severe metabolic acidosis, cardiovascular collapse, rhabdomyolysis, hyperlipidemia, renal failure, and hepatomegaly [28,33].
The incidence of PRIS is unknown, but it is probably less than 1 percent [34]. Mortality is variable but high (33 to 66 percent) [28,35,36]. Treatment involves discontinuation of the propofol infusion and supportive care [32].
Drug interactions — Central nervous system (CNS) and respiratory depressants (eg, opioid narcotics, sedatives) enhance the CNS and respiratory depressant effect of propofol. Propofol undergoes hepatic conjugation to inactive metabolites; thus, metabolically related drug interactions of major clinical importance have not been identified. Coadministration of propofol with alfentanil increases the risk of opisthotonus and/or grand mal seizures.
DEXMEDETOMIDINE — Dexmedetomidine is a highly selective, centrally acting alpha-2-agonist with anxiolytic, sedative, and some analgesic effects. According to the approved product information from the US Food and Drug Administration, dexmedetomidine is indicated for initial sedation of mechanically ventilated patients for up to 24 hours. The rationale for the 24-hour limit is that longer use may increase the risk of withdrawal effects (eg, hypertension), although these effects have not been consistently found in studies [37,38].
Efficacy — Dexmedetomidine may reduce the duration of mechanical ventilation and intensive care unit (ICU) stay when compared with traditional sedatives in the ICU, although the effect may depend upon the comparative agent [39-48]. As examples:
●A 2022 meta-analysis of 77 randomized trials reported that dexmedetomidine reduced the duration of mechanical ventilation (mean difference [MD] -1.8 h, 95% CI -2.89 to -0.71; low certainty) and ICU length of stay (MD -0.32 days, 95% CI -0.42 to -0.22; low certainty) but increased the risk of bradycardia by 6 percent and hypotension by 4 percent [48].
●When compared with midazolam, two studies suggest that dexmedetomidine decreased the duration of mechanical ventilation (SEDCOM: 3.7 versus 5.6 days; MIDEX: 5.1 versus 6.8 days, respectively) but had no effect of length of ICU stay [40-42].
●While small observational studies have shown mixed effects of dexmedetomidine in duration of mechanical ventilation [43,46,49], one large randomized study of a mixed population of critically ill patients showed no effect of dexmedetomidine (PRODEX) on the duration of mechanical ventilation when compared with propofol [40]. Similarly, another multicenter, double-blind trial of adults with sepsis (MENDS 2) also reported no difference in the number of ventilator-free days at 28 days in patients who received dexmedetomidine compared with propofol when a target goal of light sedation was set [50].
●Another trial of a mixed population of 71 ICU patients, in whom agitated delirium was prohibiting extubation, reported that compared with the addition of a placebo, the addition of dexmedetomidine to standard sedation (mostly propofol) resulted in an increase in ventilator-free hours at seven days (145 versus 128 hours) as well as reduced time to extubation (22 versus 44 hours) [51].
●A meta-analysis of four studies of a mixed population of older adults (≥60 years of age) reported that compared with propofol, dexmedetomidine did not reduce duration of mechanical ventilation (risk ratio [RR] 0.7, 95% CI 0.52-0.95) [52]. Duration of mechanical ventilation was short (<24 hours) in three of the four studies [52].
●The comparative sedative activity of dexmedetomidine versus clonidine, another systemic central alpha-2-adrenoceptor agonist, in mechanically ventilated critically ill adults is being evaluated in an ongoing randomized clinical trial [53].
Similarly, data reporting the effects of dexmedetomidine on delirium are mixed [39-41,48,50,51,54-59] but overall favor a reduction in delirium, prompting a guideline committee to issue a weak recommendation to favor dexmedetomidine as a sedative when reducing delirium was desirable by clinicians and outweighed by the undesirable effects of hypotension and bradycardia [60]:
●A 2022 meta-analysis of 77 randomized trials reported that compared with other medications, dexmedetomidine reduced the rate (or incidence) of delirium (RR 0.67, 95% CI 0.55-0.81; moderate certainty) but increased the risk of bradycardia by 6 percent and hypotension by 4 percent [48].
●In contrast, a multicenter, double-blind trial of adults with sepsis (MENDS 2) reported no difference in the number days alive without delirium or coma in patients who received dexmedetomidine compared with propofol when a target goal of light sedation was set [50].
Use of dexmedetomidine in critically ill patients does not appear to confer a mortality benefit.
●Best illustrating this was an open-label trial (SPICE III) that randomly assigned 4000 patients who were mechanically ventilated for less than 12 hours to either dexmedetomidine (as the sole sedative) or to usual care [61]. The target sedation was a Richmond Agitation and Sedation Scale (RASS) of -2 to +1 (lightly sedated to restless). Mortality at 90 days was not impacted by dexmedetomidine (29 percent in each group), and, in fact, patients receiving dexmedetomidine required supplemental sedative (eg, propofol, midazolam) to achieve the target level of sedation. In addition, bradycardia and hypotension were more common in patients receiving dexmedetomidine. This study had several limitations including a lack of blinding and a high percentage of patients who needed a deeper level of sedation (approximately 40 percent). In addition, there was no significant reduction in either ventilator- or delirium-free days in the dexmedetomidine group.
●Although dexmedetomidine has been associated with a small decrease in postoperative mortality in patients undergoing cardiac surgery, a meta-analysis of seven trials (that did not include SPICE III) reported no survival advantage associated with its use [39,55].
●In another randomized trial of 201 mechanically ventilated patients with sepsis, compared with standard sedation (with fentanyl, propofol, and/or midazolam), dexmedetomidine did not result in any mortality benefit or increase in ventilator-free days despite a higher rate of well-controlled sedation [49]. However, the study was powered to detect a 20 percent difference in mortality and, consequently, may have been unable to detect smaller differences in mortality.
●Another multicenter, double-blind trial of adults with sepsis (MENDS 2)also reported no difference in mortality at 90 days in patients who received dexmedetomidine compared with propofol when a target goal of light sedation was set [50].
Dexmedetomidine may decrease the need for alternative sedatives especially in patients withdrawing from alcohol, the details of which are discussed separately [62,63]. (See "Management of moderate and severe alcohol withdrawal syndromes", section on 'Alternative and contraindicated agents'.)
Dexmedetomidine may be more cost-effective than other sedatives. As an example, sedation with dexmedetomidine lowered ICU costs compared with standard care; cost savings were achieved primarily by reducing the duration of total ICU stay without prolonging post-ICU hospitalization [45].
Transitioning from dexmedetomidine to oral clonidine may be a safe and cost-effective way to continue sedation with a centrally acting alpha-2-agonist in patients who are hemodynamically stable and have a functional gastrointestinal tract [64]. In a single-center prospective observational pilot study, 15 of 20 patients were successfully transitioned from dexmedetomidine to clonidine [64]. Dexmedetomidine and clonidine doses were titrated during the transition period; maintenance clonidine doses ranged from 0.2 to 0.5 mg every six hours and were adjusted to achieve target sedation levels [64].
Dose regimen — An initial loading dose is typically not performed but can be administered if necessary [65]. The initial loading dose may cause transient hypotension or hypertension, depending upon whether vasodilation from activation of central alpha 2a receptors or vasoconstriction from activation of peripheral alpha 2b receptors predominates.
The usual dexmedetomidine maintenance dose is 0.2 to 1.5 mcg/kg/hour, with dose increases no more frequently than every 30 minutes [41,54]. Doses >1.5 mcg/kg/hour do not appear to add to dexmedetomidine's clinical efficacy [66]. The variability in patient response to dexmedetomidine may be due to as yet unidentified patient characteristics, pharmacokinetics, and genetic polymorphisms [67].
There are no specific guidelines for modifying the dose for older adults or patients who have renal or hepatic impairment. It is prudent to start at the low end of the dose range and titrate slowly based upon the patient's response.
The advantages, disadvantages, role, and dose regimen of dexmedetomidine are described in the table (table 1).
Adverse effects — Potential adverse events during sedation with dexmedetomidine include hypotension [40], hypertension, nausea, bradycardia [40], and atrial fibrillation. While bradycardia and hypotension can be seen with loading doses and during maintenance, hypertension can also been seen, especially during loading [68-72]. In addition, hypotension may be most commonly seen during rapid dose escalation. Refractory cardiogenic shock has also been reported [73]. Abrupt cessation should be avoided since prolonged dosing can lead to withdrawal. Thus, after cessation of dexmedetomidine, patients should be monitored for withdrawal. Case reports of fever or hyperthermia have also been reported [74-76].
Drug interactions — Although dexmedetomidine is metabolized by glucuronidation and cytochrome P450, clinically important cytochrome P450-mediated drug interactions have not been identified. Drugs that lower systemic blood pressure may enhance dexmedetomidine's hypotensive effect, while drugs that increase systemic blood pressure may enhance dexmedetomidine's hypertensive effect.
BENZODIAZEPINES — Midazolam and lorazepam are the benzodiazepines that are best suited for sedation in the intensive care unit (ICU) because they can be administered by either intermittent or continuous infusion and have a relatively short duration of effect. Intravenous diazepam is used less often to sedate patients in the ICU. It can be administered by intermittent infusion but not continuous infusion.
Mechanism — Benzodiazepines bind to specific receptors in the gamma aminobutyric acid (GABA) receptor complex, which enhances the binding of this inhibitory neurotransmitter [77]. Anxiolysis is achieved at low doses. Higher doses are associated with sedation, muscle relaxation, anterograde amnesia, anticonvulsant effects, and both respiratory and cardiovascular depression. Coadministration with an opioid analgesic may potentiate respiratory and cardiovascular depression.
Properties — The benzodiazepines are equally efficacious if they are administered in equipotent doses, but they differ in potency, rapidity of action, and duration of effect.
●Potency – A benzodiazepine's potency is determined by its binding affinity for the GABA receptor. Lorazepam has the highest binding affinity and the greatest potency. Midazolam and diazepam have progressively lower binding affinities and potencies [78].
●Rapidity of action – A benzodiazepine's rapidity of action is related to how quickly it crosses the blood-brain barrier. Midazolam and diazepam readily cross the blood-brain barrier because they are the highly lipophilic. Midazolam has an onset of action of 2 to 5 minutes following intravenous infusion, and diazepam has a nearly immediate onset of action. Lorazepam is less lipophilic and, therefore, has a slower onset of action of 5 to 20 minutes.
●Duration of effect – The duration of effect soon after initiating intermittent infusions differs from the duration of effect following repeated dosing. Initially, lipophilic benzodiazepines have a short duration of effect because there is rapid redistribution from the central nervous system to peripheral tissue sites. With repeated dosing, however, all benzodiazepines accumulate in adipose tissue. This increases the duration of effect because there is more drug that needs to be mobilized for elimination, particularly if a large cumulative dose was administered. Patients with obesity may store more drug than lean patients and are at greater risk for prolonged benzodiazepine effects. A drug's duration of effect is also influenced by the presence of active metabolites, patient factors (ie, age, body weight, hepatic function, renal function), drug interactions, and mechanism of metabolism.
Midazolam has a short duration of effect (two to four hours) when it is given short term (<48 hours) by intermittent infusion to a patient with intact hepatic function because it has rapid hepatic clearance and there is rapid redistribution to peripheral tissue sites. Midazolam may cause prolonged sedation if it is administered over a longer duration because it has a large volume of distribution, binds to peripheral tissues, and has an active metabolite (alpha-hydroxymidazolam) [79]. The active metabolite is most likely to accumulate in patients who have poor hepatic or renal function or who are receiving medications that inhibit CYP3A4 metabolism (eg, fluconazole, macrolide antibiotics, amiodarone, metronidazole).
Lorazepam has a moderate duration of effect (six to eight hours) when it is administered short-term (<48 hours) by intermittent infusion. This duration of effect reflects lorazepam's low hepatic clearance, small volume of distribution, and absence of active metabolites [80]. Lorazepam is a good choice for longer-term sedation because it has a low risk of drug interactions and its metabolism does not form active metabolites [81].
Diazepam has a short duration of effect (30 to 60 minutes) when it is administered short-term (<48 hours) by intermittent infusion. This duration of effect reflects diazepam's rapid redistribution to peripheral tissue sites and hepatic clearance. Diazepam may cause prolonged sedation with repeated dosing because it has a large volume of distribution and it has two active metabolites (desmethyldiazepam and methyloxazepam). These active metabolites are most likely to accumulate in older adults, patients who have obesity, or patients with renal or hepatic dysfunction.
Tolerance (the need for an increased dose to achieve the same effect with continued administration) occurs with all benzodiazepines. It may reflect changes in the volume of distribution or in the density, binding affinity, and/or occupancy of the benzodiazepine receptor.
Selection — The comparative advantages and disadvantages, as well as the roles of midazolam, lorazepam, and diazepam in critically ill patients, are described in the table (table 1).
Dose regimens — Typical dose regimens for midazolam, lorazepam, and diazepam are listed in the table (table 1). Dosing for patients with obesity is described separately. (See "Intensive care unit management of patients with obesity".)
Adverse effects
General — Respiratory and cardiovascular depression are well-known dose-dependent complications of benzodiazepines.
Excess sedation due to the accumulation of drug in adipose tissue can also occur as a consequence of sedation with benzodiazepines. Pharmacokinetically, this is more likely among patients who are sedated with benzodiazepines for longer than 48 hours or on continuous infusions [82].
Benzodiazepines, particularly infusions, can increase the risk for delirium in critically ill patients [54,83-87]. In an observation study of 198 mechanically ventilated patients receiving pharmacologic sedation, lorazepam was identified as an independent risk factor for delirium [83]. Delirium appears to be more common among those who receive deep sedation (even if the deep sedation is for a short duration), are older, or have dementia [84,86]. Three studies indicated that, compared with dexmedetomidine, the administration of midazolam resulted in a higher prevalence of delirium (SEDCOM: 77 versus 54 percent; MIDEX: visual assessment scale score difference of 19.7 in favor of dexmedetomidine; and MENDS trial: seven versus three days without delirium or coma) without any change in length of ICU stay [40,41] or mortality [40]. In a small, single-center unblinded study of sequential midazolam-dexmedetomidine versus sequential midazolam-propofol or midazolam alone, sequential midazolam-dexmedetomidine had a significantly lower incidence of delirium than midazolam alone (19.5 versus 43.5 percent; odds ratio 0.31, 95% CI 0.15-0.63) [88].
Patients may rarely have a paradoxical reaction to benzodiazepines. This is characterized by agitation, restlessness, and hostility [89]. Increasing the dose may worsen the agitation. The most appropriate management is to discontinue the benzodiazepine and sedate the patient with an alternative sedative. Flumazenil has been reported to reverse the paradoxical reaction [90,91].
Intravenous diazepam may increase the risk of venous thrombosis and phlebitis at the injection site [92]. The latter can cause injection site pain.
Propylene glycol toxicity — Propylene glycol is the carrier (solvent) that is used to administer intravenous lorazepam or diazepam. Infusion of either drug may be complicated by propylene glycol toxicity [93-96].
Propylene glycol toxicity is characterized by hyperosmolarity and an anion gap metabolic acidosis, which is often accompanied by acute kidney injury and can progress to multisystem organ failure, if severe [93-96]. It can occur with normal doses and renal function, but it is usually associated with doses above the recommended range of 0.1 mg/kg/hour and/or renal impairment [95,97,98]. An osmolar gap >10 mmol/L suggests that the serum propylene glycol concentration is high enough to cause toxicity [99]. Treatment consists of discontinuing the offending agent and, if severe, dialysis [96].
Propylene glycol is not the solvent for intravenous midazolam. Therefore, patients who receive intravenous midazolam are not at risk for propylene glycol toxicity.
Drug interactions — Numerous drugs used commonly in the ICU may interact with benzodiazepines. Some increase the benzodiazepine effect, while others decrease the effect.
Central nervous system (CNS) and respiratory depressants (eg, opioids) enhance the CNS and respiratory depressant effect of benzodiazepines. Conversely, CNS stimulants (eg, methylphenidate) decrease the CNS depressant effect of benzodiazepines. Many other drug interactions are related to the metabolism of benzodiazepines via the cytochrome P450 system:
●Midazolam is susceptible to interactions with drugs that either inhibit or induce CYP3A4, since CYP3A4 hydroxylates midazolam to two active metabolites, 1-hydroxy-midazolam and 3-hydroxy-midazolam. As examples, the azole antifungals (eg, fluconazole, itraconazole, ketoconazole, voriconazole) and the macrolides and related antibiotics (eg, clarithromycin, erythromycin, telithromycin) may prolong midazolam activity by inhibiting CYP3A4. Conversely, carbamazepine may decrease midazolam activity by inducing CYP3A4. Rifamycins (eg, rifampin, rifabutin) may decrease midazolam activity by increasing cytochrome P450-mediated oxidative metabolism. Hydantoins (eg, phenytoin, fosphenytoin) may increase midazolam clearance and midazolam inhibits hydantoin clearance; the latter leads to increased serum concentrations of the hydantoins.
●Diazepam is susceptible to interactions with many of the same drugs that inhibit or induce cytochrome P450 because it is hydroxylated by CYP3A4 to temazepam and N-demethylated by CYP3A4 and CYP2C19 to desmethyldiazepam. Temazepam and desmethyldiazepam are active metabolites that are subsequently metabolized to another active metabolite, oxazepam.
●Lorazepam does not interact with drugs that inhibit or induce cytochrome P450, because it undergoes extensive glucuronidation in the liver to an inactive 3-O-phenolic metabolite.
KETAMINE — Ketamine is rarely used and is not approved for use in the adult intensive care unit (ICU) population. Unlabeled uses include procedural sedation/analgesia, treatment of extreme agitation, and as an adjunct to opioid analgesia [7].
Ketamine is an intravenous anesthetic agent with analgesic and bronchodilator properties in subanesthetic doses. Ketamine noncompetitively blocks glutamate N-methyl-D-aspartate (NMDA) receptors within sensory nerve endings; other pharmacologic actions at subanesthetic doses have been identified, including opioid and muscarinic agonist activities, and nicotinic receptor blockade [100].
Due to its lipid solubility, an intravenous bolus dose of ketamine is active within one minute with a duration of action of 10 to 15 minutes. Ketamine is metabolized in the liver through several cytochrome systems to several metabolites including the weakly active metabolite norketamine that are cleared by the kidneys.
Ketamine produces a "dissociated anesthesia," wherein patients remain conscious with spontaneous breathing and intact brain stem reflexes. By stimulating the sympathetic nervous system, there is less cardiovascular depression; this preserves and sometimes increases blood pressure, making it an attractive agent for use in patients in shock or with frank hypotension. Ketamine also has mild bronchodilatory activity such that it has been anecdotally used for sedation in patients with status asthmaticus.
The use of ketamine is limited by its psychoactive effects (vivid hallucinations, confusion, and delirium). Hallucinations or delirium may also occur during recovery from ketamine. Ketamine is contraindicated in patients with known hypersensitivity to it and in patients at risk from potentially significant elevations in blood pressure. Other significant adverse reactions include excessive salivation and respiratory and cardiac depression.
Small randomized studies of patients with burns suggest that during painful procedures oral ketamine provides better analgesia than dexmedetomidine or the combination of midazolam, acetaminophen, and codeine (eg, dressing changes) [101,102]. Another review suggested a reduction in opioid consumption postoperatively [103]. Data from well-designed randomized controlled trials are needed to determine a clear role of ketamine in procedural sedation/analgesia and ICU analgesia. (See "Pain control in the critically ill adult patient", section on 'Ketamine' and "Management of burn wound pain and itching", section on 'Nonopioid analgesics'.)
ANTIPSYCHOTICS — Antipsychotics can be used in the intensive care unit (ICU) for the treatment of delirium. Haloperidol can be administered intravenously, has a mild sedative effect, and has relatively low cardiorespiratory depressive effects. Despite its widespread use for the treatment of delirium in the ICU, there is no published evidence supporting a reduction in the duration of mechanical ventilation or duration of delirium by haloperidol and guidelines make no recommendation favoring its use over other antipsychotics for the management of delirium in adult ICU patients [7].
Atypical antipsychotics (quetiapine, olanzapine, risperidone, ziprasidone) have also been used in adult ICU patients to treat delirium. While there is some evidence that the oral atypical antipsychotics improve delirium in critically ill patients [104,105], there is a paucity of studies that examine outcome or compare the efficacy and safety of oral atypical antipsychotics to haloperidol and to one another. The few studies that exist suggest that the efficacy and safety of oral atypical antipsychotics may be similar to that of haloperidol [104,106,107]. In a retrospective review of 156 adult patients admitted to an ICU and treated with an atypical antipsychotic for management of delirium, 31 percent developed QTc prolongation; 24 percent had a QTc greater than 500 msec [108]. (See 'Adverse effects' above.)
Although haloperidol and atypical antipsychotics can be used, further studies are necessary to validate the role of haloperidol or oral atypical antipsychotics, such as quetiapine, as effective therapies to reduce the duration of mechanical ventilation in adult ICU patients.
Among ventilated patients, haloperidol does not appear to prevent or decrease the duration of delirium or mortality [109-113].
●One randomized trial of 1789 critically ill patients at risk of delirium (defined as an anticipated intensive care unit stay of at least two days) reported that compared with placebo, 2 mg of haloperidol administered intravenously three times a day had no impact on the incidence of delirium, survival, duration of mechanical ventilation, or length of stay [111]. Although outcomes may have been affected by factors including the high prevalence of nonpharmacologic interventions that were integrated into daily ICU care and heterogeneity of background treatment across multiple sites, the consistent lack of impact of haloperidol on any of the 16 outcomes studied suggests that the lack of benefit is real. This study justifies not using haloperidol prophylactically in the ICU for the prevention of delirium.
●In another randomized trial of 566 ICU patients with hyperactive or hypoactive delirium, compared with placebo, there was no difference in the number of days alive without delirium or coma when patients were treated with haloperidol or the atypical antipsychotic, ziprasidone [112]. There was also no difference in 30- or 90-day mortality, duration of mechanical ventilation, or time to ICU or hospital discharge.
●In a randomized trial of 1000 mechanically ventilated patients, treatment of delirium with haloperidol did not alter the number of days alive outside the hospital at 90 days compared with placebo [113]. The overall adverse event rate was low and was no different between the groups. However, this trial did not collect data on other sedatives used, and the number of patients with hypoactive delirium (as opposed to hyperactive delirium) may have been under enrolled. Nonetheless, these findings are in keeping with other studies and suggest no mortality benefit when haloperidol is used to treat delirium in critically ill patients [112].
Mechanism — Haloperidol and the other neuroleptics antagonize dopamine and other neurotransmitters. However, their precise mechanism of action remains unknown.
Properties — Haloperidol causes dose-dependent sedation. It tends to be less sedating and has less anticholinergic activity than other neuroleptics.
●Rapidity of onset – Haloperidol has an onset of action 5 to 20 minutes after intravenous infusion.
●Duration of effect – Haloperidol's duration of effect varies and depends upon the cumulative dose. Generally speaking, redosing may be needed 4 to 12 hours after symptoms have been controlled with the initial doses.
Haloperidol is highly protein bound; has a large volume of distribution; and is metabolized hepatically by CYP3A4, CYP2D6, and glucuronidation; the hydroxymetabolite (reduced haloperidol) is active [114]. The pyridinium metabolite, a structural analogue of a known neurotoxin, may be neurotoxic [115].
Dose regimens — The administration of haloperidol intravenously is common, but it has not been approved by the US Food and Drug Administration (FDA). The reason that it has not been approved is that intravenous administration may have a greater risk of serious adverse events than oral or intramuscular administration, most notably torsades de pointes and sudden cardiac death [116].
Numerous dosing regimens have been used for the treatment of delirium in adult ICU patients, none of which have been validated. Continuous infusions (table 1) are rarely indicated, but studies suggest that they are probably safe and effective [117-119]. Most commonly used regimens are:
●2.5 to 5 mg intravenous bolus doses administered every six hours, as needed (table 1) [120].
●An initial dose determined by the severity of the agitation. Examples include a 0.5 to 2 mg intravenous bolus dose for mild agitation, a 2 to 5 mg intravenous bolus dose for moderate agitation, and a 10 to 20 mg intravenous bolus dose for severe agitation. Following the initial dose, some clinicians give repeat doses as frequently as every 15 to 30 minutes in patients with severe agitation until the desired level of sedation is achieved.
●A continuous haloperidol infusion beginning at a dose of 10 mg/h and then increasing 5 mg/hr every thirty minutes as needed until calm is achieved [117-119].
Once calm is restored, a maintenance dose is desirable unless it appears that the delirium may quickly resolve. A reasonable approach to maintenance dosing is to administer 25 percent of the total loading dose every six hours (table 1).
The safe maximum daily dose of haloperidol is not known. There are case reports of doses as high as 945 mg/day [117,121]. Doses greater than 200 mg/day have been safely administered for up to 15 consecutive days [117,122].
Adverse effects — Haloperidol-associated polymorphic ventricular tachycardia (including torsades de pointes) is an uncommon but severe adverse reaction [123,124]. It is primarily associated with intermittent high dose intravenous administration and prolonged QTc interval. When intermittent haloperidol infusions are used, the QT interval should be monitored every shift (ie, every 8 to 12 hours) and haloperidol should not be given if the corrected QT interval exceeds 500 msec. (See "Acquired long QT syndrome: Definitions, pathophysiology, and causes".)
Other potential side effects of haloperidol include acute dystonic reactions, parkinsonism, tardive dyskinesia, akathisia, and neuroleptic malignant syndrome. For unclear reasons, extrapyramidal side effects are less common among patients receiving intravenous haloperidol than among those receiving oral haloperidol [125]. (See "Treatment of dystonia in children and adults".)
Older adults with dementia may have an increased risk of cardiovascular-related death (due to heart failure or sudden cardiac death) or infection-related death (due to pneumonia) when treated with atypical or typical antipsychotics. This was illustrated by a meta-analysis of 15 randomized trials (5204 patients) that found that older patients with dementia who were treated with atypical antipsychotics had an increased risk of death compared with those who received placebo (3.5 versus 2.2 percent; odds ratio 1.54, 95% CI 1.06-2.23) [126]. These findings were supported by two subsequent studies [127,128]. The FDA has since issued a black box warning for the atypical antipsychotics. Prescribing information for all antipsychotics now includes the black box warning. Antipsychotics (typical and atypical) should be used with caution in all older adults with dementia, including the critically ill.
Drug interactions — Haloperidol interacts with numerous drugs that are common in the ICU. As an example, drugs with central nervous system (CNS) depressant effects (eg, opioids, sedatives) may enhance the CNS depressant effect of haloperidol.
Other drug interactions are related to the metabolism of haloperidol via the CYP3A4 and CYP2D6 pathways. These include azole antifungals and carbamazepine:
●Systemic azole antifungals (eg, fluconazole, itraconazole, posaconazole, voriconazole), HIV and HCV protease inhibitors, and cyclosporine, inhibit CYP3A4, which prolongs haloperidol activity.
●Carbamazepine increases CYP3A4 metabolism, which decreases haloperidol activity. This effect is enhanced because haloperidol increases carbamazepine activity by inhibiting carbamazepine metabolism.
The rifamycins (eg, rifampin, rifabutin) also induce cytochrome P450-mediated oxidative metabolism and decrease haloperidol activity. Anticholinergics (eg, atropine, glycopyrrolate) increase haloperidol clearance via an unknown mechanism.
When haloperidol is used, other drugs that enhance the QTc-prolonging effect of haloperidol (eg, amiodarone, dronedarone, ranolazine, methadone, high-dose ondansetron, domperidone, erythromycin, fluoroquinolone antibiotics, tricyclic antidepressants, posaconazole, voriconazole) should be avoided. Additional agents are listed in the table (table 2). Metoclopramide may increase haloperidol toxicity and should also not be given concurrently.
BARBITURATES — Thiopental (Pentothal) and methohexital (Brevital) are barbiturates that are occasionally used to sedate critically ill patients. Barbiturates produce sedation by binding to the gamma aminobutyric acid (GABA)-receptor complex via a different receptor from benzodiazepines. They commonly cause hypotension and may produce profound cardiovascular and respiratory depression. As a result, the use of barbiturates should be limited to patients not tolerating or responding to other agents. Other undesirable characteristics of barbiturates include prolonged elimination half-lives, induction of the cytochrome p-450 enzyme system, and accumulation of drugs in renal and hepatic dysfunction. Thiopental is no longer manufactured in the United States or Canada.
SEVOFLURANE — Sevoflurane, a polyfluorinated methyl-isopropyl compound, is a volatile inhalational anesthetic that is being evaluated as a potential sedative agent for intensive care unit (ICU) patients. Ninety-five to 98 percent of sevoflurane is eliminated through the lungs, while the remaining 2 to 5 percent undergoes rapid hepatic metabolism to inorganic fluoride and hexafluoroisopropanol (HFIP) [129]. HFIP in the blood is conjugated by glucuronic acid and then secreted by the kidney.
Potential advantages of sevoflurane as a sedating agent for critically ill ICU patients include the short duration of action and rapid elimination [130]. Potential disadvantages include fluoride accumulation with prolonged use (especially in patients with impaired renal function) and malignant hyperthermia [130]. Additionally, sevoflurane undergoes degradation on contact with alkaline carbon dioxide absorbents used to remove carbon dioxide from the circuit, to a potentially nephrotoxic product (trifluoromethyl vinyl ether; Compound A) [129]. Administration of volatile anesthetic agents in the ICU is not a standard practice; investments in technology and clinician education will be required before sevoflurane can be used as a routine sedative in critically ill patients. (See "Malignant hyperthermia: Diagnosis and management of acute crisis" and "Susceptibility to malignant hyperthermia: Evaluation and management".)
A trial randomly assigned 60 mechanically ventilated ICU patients to sedation with sevoflurane, propofol, or midazolam [131]. All of the patients received remifentanil for analgesia, most were relatively young trauma patients, and 47 completed the trial. The median duration of sedation in the sevoflurane group was 50 hours (range 39 to 71 hours). Wake-up time and time to extubation were significantly shorter in the sevoflurane group compared with the propofol or midazolam groups. The study had too few events to conclusively detect differences in ICU length of stay or mortality.
BACLOFEN — Baclofen is a gamma-aminobutyric acid type B receptor agonist that at high levels can reduce consciousness. Its role as a sedative in the intensive care unit (ICU) is unclear, but previous data suggest that baclofen may mitigate alcohol craving in patients with alcohol use disorder [132,133]. One randomized trial used high-dose oral baclofen (50 to 150 mg per day) for the management of agitation in 314 mechanically ventilated patients who had unhealthy alcohol use (defined as consumption of more than 14 units per week for males and more than 7 units per week for females and males >65 years) [134]. Baclofen resulted in a reduction in the proportion of patients who experienced at least one agitation-related event (19.7 versus 29.7 percent; eg, self-extubation, removal of venous catheters) but did not impact the 28-day ICU mortality when compared with placebo. In addition, baclofen was associated with a longer duration of mechanical ventilation and ICU length of stay. The most common adverse effect was delayed awakening; others included stroke, seizure, and bradycardia. Several flaws in study design including the lack of standardized baclofen dosing, inaccurate assessment of alcohol consumption, and lack of parenteral formulation limit interpretation and generalization of this study. Further studies are needed before baclofen can be recommended for use in individuals with agitation from unhealthy alcohol use.
CHOICE OF AGENT — No sedative-analgesic agent is sufficiently superior to other agents to warrant its use in all clinical situations. As a result, selection of an agent must be individualized according to patient characteristics and the clinical situation. The etiology of the distress, expected duration of therapy, potential interactions with other drugs, desired depth of sedation, and pharmacokinetic-modifying factors are important considerations whenever selecting an agent. The selection of sedative analgesics in the critically ill patient is discussed separately. (See "Sedative-analgesic medications in critically ill adults: Selection, initiation, maintenance, and withdrawal", section on 'Selection of an agent'.)
COMBINATION THERAPY — Many clinicians believe that combination therapy using different types of sedatives is best. The rationale is that agitation is often multifactorial and each potential cause deserves a targeted therapy. This notion was supported by a trial that randomly assigned 30 mechanically ventilated patients to receive midazolam alone or midazolam plus fentanyl [135]. Midazolam plus fentanyl maintained sedation level goals better, decreased the dose of the primary agent, added analgesia, and did not appreciably increase the likelihood of prolonged sedation. There were no differences in hemodynamic or respiratory adverse effects. Two patients treated with the combination regimen developed ileus, compared with none with midazolam alone. (See "Sedative-analgesic medications in critically ill adults: Selection, initiation, maintenance, and withdrawal".)
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 due to pain, anxiety, dyspnea, or delirium is common among critically ill patients. Distress may cause ventilator asynchrony and increase sympathetic tone, which may have untoward clinical effects. (See 'Introduction' above.)
●Classes of agents – Classes of sedative-analgesic medications used to treat distress include opioids, propofol, dexmedetomidine, benzodiazepines, neuroleptics, and, rarely, barbiturates. Pilot studies of the propofol prodrug fospropofol, the volatile anesthetic sevoflurane, and baclofen have been reported. (See 'Benzodiazepines' above and 'Analgesics' above and 'Antipsychotics' above and 'Propofol' above and 'Dexmedetomidine' above and 'Barbiturates' above and 'Sevoflurane' above.)
●Agent selection – No sedative-analgesic agent is sufficiently superior to other agents to warrant its use in all clinical situations. As a result, selection of an agent must be individualized according to patient characteristics and the clinical situation. The etiology of the distress, expected duration of therapy, potential interactions with other drugs, desired depth of sedation, and pharmacokinetic-modifying factors are important considerations whenever selecting an agent. (See "Sedative-analgesic medications in critically ill adults: Selection, initiation, maintenance, and withdrawal", section on 'Selection of an agent'.)
●Adverse effects – Dose-dependent respiratory and cardiovascular depression are common among sedative-analgesic agents. Most classes of sedative analgesics also have unique adverse effects. Examples include propylene glycol toxicity (eg, from lorazepam or diazepam) and the propofol infusion syndrome. (See 'Propylene glycol toxicity' above and 'Propofol-related infusion syndrome' above.)