Opioids are potent, predictable, and reversible analgesics but are poor amnestic agents. When analgesic doses of opioids are used alone, few hemodynamic or respiratory effects are observed; however, when large “anesthetic-range” doses (often ten times the analgesic dose) are used or when narcotics are combined with neuromuscular blockers or other sedatives, the risk of cardiopulmonary instability is magnified. Opioid complications are minimized by (i) ensuring intravascular volume is adequate, (ii) using the lowest effective dose, and (iii) slow administration. The histamine-releasing (vasodilating) potential of opioids is minimal unless rapid, large intravenous doses are given (meperidine and morphine are the most common offenders). Although rarely necessary, histamine effects can be attenuated by pretreatment with H₁ and H₂ blockers. A feature common to all narcotics is blunting of the hypercapnic and hypoxic respiratory drives. Although often considered a liability, lessening respiratory drive often benefits mechanically ventilated patients by reducing their sense of dyspnea and minute ventilation requirement and can therefore decrease the tendency for breath stacking. Blunting of respiratory drive can also provide great comfort for dyspneic patients at the end of life.

Unfortunately, opioids and their breakdown products accumulate in patients with hepatic and/or renal failure, especially in those receiving prolonged treatment. Hepatic biotransformation typically precedes renal excretion of drugs and their metabolites. Initially, the synthetic, highly lipid-soluble opioids (e.g., fentanyl, sufentanil, alfentanil) have their actions terminated by redistribution, not metabolism. With chronic use, however, patients become “saturated” with the drug, requiring metabolism for termination of effect. Another synthetic opioid, remifentanil, has its effects terminated within minutes by rapid plasma esterase metabolism. Opioids, especially in high doses, may complicate attempts at enteral feeding by reducing GI motility and inciting nausea. Although still preliminary, recent reports suggest subcutaneous administration of the μ receptor antagonist methylnaltrexone may attenuate this effect. Rarely, biliary spasm may be precipitated by opiate use, but for the patient with biliary colic the analgesic benefits far outweigh the theoretical risks. Regional and epidural blocks, patient-controlled analgesia, multimodal analgesia using nonopioid analgesics (e.g., ibuprofen), and addition of anxiolytic agents which potentiate analgesics (e.g., benzodiazepines) are significant advances in pain control, providing superior pain relief with a lower total narcotic dose and lower risk of oversedation. Reluctance to use narcotics for long-term pain relief for fear of dependence or addiction is unfounded; addiction is rare among patients with real pain who lack a history of substance dependence. However, there is a growing appreciation that withdrawal does occur in long-term recipients of high doses of opioids who have their therapy abruptly discontinued. Withdrawal syndromes in the ICU may masquerade as infection because they manifest as fever, tachycardia, tachypnea, and confusion.

Morphine is an inexpensive drug with a rapid onset of action and a 1- to 3-h half-life. Intermittent intravenous (IV) doses of 2 to 10 mg or 1 to 3 mg/h (0.03 to 0.15 mL/kg/h) by constant infusion usually are adequate for relief of moderate to severe pain in the average adult. Administration rates less than 10 mg/min minimize the risk of hypotension. Although morphine is an excellent analgesic, high doses or the addition of a benzodiazepine are usually required to produce unconsciousness. Morphine’s action is prolonged by both renal and hepatic failure because a fraction of each morphine dose is directly excreted unchanged by the kidney, but most is metabolized hepatically before renal excretion. One advantage of morphine over synthetic opioids, especially in the oliguric patient, is its high water solubility, permitting analgesia to be administered in a minimum volume. Hydromorphone, is an even more potent (approx. 10-fold) semisynthetic opiod derived from morphine. Its half-life of 2 to 3 h provides a similar or slightly longer analgesic effect than morphine. However, because the half-life of hydromorphone can increase 20-fold in patients with renal failure, extreme caution must be exercised in this population.

Fentanyl is a potent, highly lipid-soluble, synthetic opioid possessing a very rapid onset and brief duration of action—at least with initial use. With repeated injections or when given by continuous
infusion, large stores of drug may accumulate in lipophilic tissues that then must be metabolized to terminate the drug’s action. Because of this accumulation, fentanyl’s effective half-life after days of use may exceed that of morphine. Initial analgesic doses of 0.5 to 10 μg/kg IV may be titrated upward as necessary. Because fentanyl cannot be prepared in highly concentrated aqueous solutions, it is one of the few medications that can present a “volume” challenge. With chronic use, it is common to req uire several hundred milliliter of fluid each day to administer the analgesic dose. Fentanyl’s perceived chief advantage is its minimal hemodynamic effect for a given level of analgesia. Very rarely, fentanyl causes seizures or a bizarre syndrome of chest wall rigidity (most common when large IV doses are given rapidly to elderly patients). Thoracic rigidity may be so severe that intubation, neuromuscular paralysis, and mechanical ventilation are necessary. Transdermal patches bypass the substantial first-pass hepatic clearance seen with IV dosing, offering a useful alternative method of administration. Unfortunately, the skin slows diffusion, resulting in a long lag time between patch application and effective analgesia, and removal of the patch fails to rapidly terminate the drug’s effect because the skin serves as a “reservoir.” For patients with modest analgesic requirements and good perfusion, the patch delivery system is especially worth considering. Unfortunately, a very high incidence of nausea has been reported with transdermal dosing. As an aside, patients presenting with a clinical syndrome of opioid overdose should be carefully examined to make sure they do not have one or more fentanyl patches affixed.

Alfentanil is a short-acting, lipid-soluble, synthetic opioid that is more potent than morphine but less so than fentanyl. Consciousness returns rapidly after high doses, making this drug useful for brief but painful procedures. Absence of active metabolites results in minimal drug accumulation unless hepatic failure is present. Sufentanil is perhaps 1,000 times as potent as morphine but, except for a potentially smaller volume requirement, does not offer any significant advantages over fentanyl. Remifentanil, a very potent, short-acting, synthetic opioid, is metabolized by plasma esterases, thus its metabolism is not altered by hepatic or renal failure. Alfentanil, sufentanil, and remifentanil are expensive options without unique therapeutic effect.

Meperidine should be avoided. It is a poor pain reliever; offers no substantial advantage over other opioids, and its myocardial depressant effect and vagolytic and histamine-releasing tendencies often cause tachycardia and hypotension. The perceived superiority of meperidine for patients with ureteral or biliary colic is unfounded. Meperidine’s major metabolite, normeperidine, is active, accumulates in renal failure, and causes seizures when present in high concentrations.

In the event of opioid overdose, the antagonist naloxone promptly reverses excessive sedation. Intravenous doses of 0.4 to 2 mg are usually sufficient, at least transiently, although doses as high as 10 mg are occasionally required. Naloxone given in repeated smaller doses or by slow intravenous infusion can undo the respiratory depressant effects of narcotics without reversing the analgesia. Low-dose naloxone is particularly useful for chronic opioid users who unintentionally develop excessive sedation. Naloxone’s duration of action is not as long as many commonly used opioids, necessitating close observation and sometimes-repeated administration to prevent recurrence of sedation.

Nonsteroidal anti-inflammatory agents (NSAIDs) are often avoided in the ICU because of their antiplatelet activity and reputation to cause bleeding and renal insufficiency. The frequency and severity of these adverse events are overestimated, and unfortunately drugs with more serious side effects are often chosen as alternatives. Oddly, NSAIDs are often blamed when renal insufficiency develops in patients with multiple potential causes of kidney injury (e.g., low cardiac output, hypotension, contrast exposure, high-dose vasoconstrictors, angiotensin converting enzyme [ACE] inhibitors). Whereas the cyclooxygenase and platelet-inhibiting activities make NSAIDs less than ideal choices for patients with impaired renal function, coagulopathy or active bleeding, antiplatelet activity may be advantageous for patients with ischemia. These potent, inexpensive compounds (e.g., aspirin, ibuprofen, and COX-2 inhibitors) often suffice for pain relief and act synergistically with narcotics. (In studies of postoperative pain, opioid doses may be reduced by 1/4 to 1/3). No credible data support enhanced GI safety of selective COX-2 inhibitors in the ICU population. The major drawback is that most NSAIDs can only be given via the GI tract. (Ibuprofen is well absorbed from the rectum, and
a liquid formulation can be used in patients unable to tolerate gastric administration.) Currently, only ibuprofen and ketorolac are offered in parenteral form.

Benzodiazepines are sedative, anxiolytic amnestics with a wide therapeutic margin. Because pain and anxiety are synergistic and often indistinguishable, benzodiazepines can reduce analgesic needs even though they have no intrinsic analgesic properties. Benzodiazepines also induce amnesia and provide anticonvulsant and muscle-relaxant properties. Although not as potent as barbiturates or propofol for the purpose, benzodiazepines reduce cerebral O₂ consumption, intracranial pressure, and cerebral blood flow. Although they may induce unconsciousness, this state is not normal slumber. Sleep fragmentation occurs without the full range and depth of normal sleep stages. When sleep is the primary objective, a hypnotic (e.g., zolpidem or triazolam) are better choices than benzodiazepines intended for sedation (e.g., midazolam, lorazepam, or diazepam). Benzodiazepines are associated with the development of delirium in the critically ill and paradoxically, can excite patients by disinhibiting normal social behavioral control. Similarly, amnestic and dissociative effects may linger after consciousness returns, resulting in an agitated, confused state that prompts additional doses of the very medication that precipitated it—a potentially vicious cycle. Droperidol and its parent drug, haloperidol, are effective agents to break this cycle.

Unless used in large doses or combined with narcotics, propofol, or neuromuscular blockers, benzodiazepines have few cardiovascular effects. Mild tachycardia and minimal reductions in blood pressure are most commonly observed in elderly or dehydrated patients, patients using b-blockers, and patients with underlying cardiac disease. Benzodiazepines cause mild dose-dependent respiratory depression but rarely cause apnea. (Apnea is most common after rapid administration of large IV doses to chronically ill patients or elderly patients or patients receiving concomitant narcotics.) The highly lipid-soluble benzodiazepines (e.g., midazolam, diazepam) accumulate in fat after repeated or prolonged use, resulting in delayed recovery. The avid protein binding of benzodiazepines leads to frequent interactions with other protein-bound drugs and exposes hypoproteinemic patients to high concentrations of free (active) drug. Most benzodiazepines require hepatic metabolism and/or excretion; therefore, liver disease can prolong the action of these drugs (lorazepam and oxazepam are least subject to this effect). Conversely, patients with induced liver enzymes (e.g., alcoholics, barbiturate users) may require enormous doses for effect. Currently, all the commonly used benzodiazepines are inexpensive and comparably priced based on hourly use. The properties of the three most frequently used parenteral benzodiazepines are contrasted in Table 17-2.