Management of Opioid-Induced Adverse Effects

Opioid Antagonists for Bowel Dysfunction

Although receptors in the central nervous system (CNS) are involved in the pathophysiology of opioid-induced constipation, the effect appears to be mediated predominantly by GI mu opioid receptors (Thomas, Karver, Cooney, et al., 2008). This observation led to a search for a peripherally acting mu opioid antagonist that could specifically reverse opioid-induced bowel dysfunction without reversing analgesia (Thomas, 2008).

When taken orally, the opioid antagonists naloxone, naltrexone, and nalmefene are absorbed systemically, and although they can effectively reverse bowel dysfunction, they can cross the blood-brain barrier, reverse central opioid receptors and analgesia, and produce withdrawal symptoms (Becker, Galandi, Blum, 2007; Goodheart, Leavitt, 2006; Liu, Wittbrodt, 2002; Thomas, 2008). This outcome is least likely to occur with naloxone, which has a very limited oral bioavailability (around 3%). Although there is a risk of systemic absorption with associated return of pain and withdrawal, the risk with this drug is low, and oral naloxone (20 to 40 mg) has been used clinically for refractory constipation (Meissner, Leyndecker, Mueller-Lissner, et al., 2009).

In contrast to these other antagonists, methylnaltrexone (Relistor) and alvimopan (Entereg) have the potential to block opioid actions mediated by peripheral opioid receptors while sparing actions mediated by opioid receptors in the CNS (Portenoy, Thomas, Moehl Boatwright, et al., 2008). Methylnaltrexone has approval in the United States for the treatment of opioid-induced constipation in patients with advanced illness (Wyeth Pharmaceuticals, 2009), and alvimopan is approved for acceleration of the time to upper and lower GI recovery following partial large or small bowel resection with primary anastomosis (GlaxoSmithKline, 2009). Methylnaltrexone is given subcutaneously (8 mg for patients weighing 38 to 62  kg every other day and no more than one dose in a 24-hour period), and alvimopan is taken orally (12 mg 30 minutes to 5 hours prior to surgery, then 12 mg twice daily for up to 7 days for a maximum of 15 doses). Methylnaltrexone is not approved for IV use but has been used by this route for treatment of nausea, pruritus, and urinary retention (Gold Standard Clinical Pharmacology, 2009).

A systematic review of 20 studies that were conducted on the use of methylnaltrexone and alvimopan for treatment of constipation concluded that further research with larger numbers of patients and varying types of pain are required (Becker, Galandi, Blum, 2007). Another analysis of 22 studies concluded that there is insufficient evidence for the safety and efficacy of naloxone or nalbuphine for the treatment of opioid-induced bowel dysfunction and that further research is needed to fully assess the role of alvimopan and methylnaltrexone in therapy (McNicol, Boyce, Schumann, et al., 2008). Following is a discussion of some of the research on methylnaltrexone. See pp. 491-493, Postoperative Ileus, for more on alvimopan.

One of the first studies conducted on methylnaltrexone was a double-blind study of 22 adults who were enrolled in a methadone maintenance program and had methadone-induced constipation (Yuan, Foss, O’Connor, et al., 2000). The subjects were randomized to receive either IV methylnaltrexone (up to 0.365  ) or placebo infusion over 9 minutes. All of the subjects who received methylnaltrexone had abdominal cramping followed by laxation on day 1 or day 2; none of those who received placebo had these effects. The effects of methadone maintenance were not reversed in this study.

A 2-week double-blind trial randomized 133 patients with incurable cancer or other end-stage disease who had been taking stable doses of opioid analgesics for 2 or more weeks to receive either SC methylnaltrexone (0.15  ) or placebo (Thomas, Karver, Cooney, et al., 2008). The median dose of morphine equivalent the patients were taking was 100 mg and 150 mg in the methylnaltrexone and placebo groups, respectively, and the patients were constipated at baseline. Laxation occurred within 4 hours of the first study dose in 48% of those who received methylnaltrexone compared with 15% of those who received placebo, and 52% had laxation without a rescue laxative within 4 hours after two or more of the first four study doses compared with 8% in the placebo group. Adverse effects were mild or moderate (e.g., abdominal pain, flatulence, nausea, increased body temperature, and dizziness) in 8% and 13% of those in the methylnaltrexone and placebo groups, respectively. Life-threatening adverse events were assessed as related to the primary illness. Eighty-nine of the patients in this study entered a 3-month, open-label extension study in which methylnaltrexone was administered to all of them. The response rate in those who had received methylnaltrexone and placebo in the double-blind phase was 45% to 58% and 48% to 52%, respectively. Similar adverse effects occurred in the open-label phase as in the double-blind phase. Analgesia was maintained throughout both phases of this study.

Methylnaltrexone was studied in a randomized, parallel-group, repeated dose, dose-ranging trial that included a one-week double-blind phase followed by an open-label phase for up to 3 weeks (Portenoy, Thomas, Moehl Boatwright, et al., 2008). The patients (N = 33) in this study had terminal or end-stage diseases and were receiving palliative care and long-term opioid therapy with stable doses for at least 2 weeks (mean and median opioid morphine equivalent dose = 289.9 mg and 180 mg, respectively) and reported ongoing constipation. They were randomized to receive 1, 5, 12.5, or 20 mg of SC methylnaltrexone. Doses between 5 mg and 20 mg (0.05 to 0.5  ) induced a bowel movement within 4 hours of drug administration significantly more often than a dose of 1 mg (less than 0.05  ), and there was no dose-response relationship above 5 mg/day. Approximately 50% of the patients responded with doses 5 mg or more within 4 hours and maintained favorable effects with repeated doses. These doses produced effective and rapid relief of constipation without producing pain flare or opioid withdrawal symptoms. All of the patients in this study experienced at least one adverse effect related to treatment; however, most were mild and not related to the dose of methylnaltrexone; abdominal pain was the most common. More recent research (N = 52) showed methylnaltrexone (0.15  ) given SC every other day for 2 weeks to patients with advanced illness resulted in a higher percentage of patient-rated improvements in bowel status, prompt and predictable laxation, and less use of other laxation techniques (e.g., laxatives and enemas) compared with placebo (Chamberlain, Cross, Winston, et al., 2009) (see Table 19-1).

Opioid Antagonists for Management of Ileus

A major advance in the management of postoperative ileus is the approval of alvimopan (Entereg) for the acceleration of the time to upper and lower GI recovery following partial large or small bowel resection with primary anastomosis (GlaxoSmithKline, 2009). Alvimopan is a synthetic peripherally acting mu opioid receptor antagonist taken orally (12 mg 30 minutes to 5 hours prior to surgery, then 12 mg twice daily for up to 7 days for a maximum of 15 doses) (Kraft, 2007). Randomized controlled trials have shown that the drug can reduce ileus, shorten hospital length of stay, and is well tolerated with adverse effects similar to placebo after major abdominal surgery (Becker, Blum, 2009; Leslie, 2005; Neary, Delaney, 2005; Sinatra, 2006; Viscusi, Gan, Leslie, et al., 2009; Viscusi, Goldstein, Witkowski, et al., 2006). Methylnaltrexone, another peripherally acting mu opioid antagonist, has also been shown to reduce ileus (Viscusi, Gan, Leslie, et al., 2009) but is approved and used most often for the treatment of opioid-induced constipation in patients with advanced illness (see previous discussion earlier in the chapter). The opioid antagonists naloxone and nalmefene are not selective for the mu opioid receptors in the GI tract and should not be used for the prevention or treatment of ileus (Kraft, 2007).

In summary, postoperative ileus has multiple underlying mechanisms and is influenced by a number of factors. Management requires the implementation of a multimodal approach that focuses on prevention. Strategies include continuous thoracic epidural analgesia, opioid-sparing analgesic techniques, peripheral mu opioid antagonists, laparoscopic surgical techniques, and avoidance of routine NG tube, fluid excess, and immobility (see Table 19-1).

• Identify patients at high risk for PONV (see Box 19-1). There is no consensus on how many risk factors a patient must have to warrant the designation of high risk (Apfel, Kranke, Eberhart, et al., 2002; Gan, Meyer, Apfel, et al., 2003; van den Bosch, Kalkman, Vergouwe, et al., 2005). A simple scoring system (the Apfel risk score) developed by Apfel and colleagues (1999) is widely used and has been shown to be reliable and valid for this purpose. It is based on four predictors: female gender, prior history of motion sickness or PONV, nonsmoking status, and the use of postoperative opioids. If no or only one factor is present, the incidence of PONV varies from 10% to 21%. If two or more are present, the risk rises to 39% to 78% (Apfel, Laara, Koivuranta, et al., 1999).

• Reduce baseline risk factors, e.g., implement multimodal analgesic strategies to treat postoperative pain so that no opioid or the lowest effective dose of opioid can be given. This may also include the use of effective nonpharmacologic strategies such as relaxation techniques, acupuncture, and acupressure (Lee, Done, 1999; Nunley, Wakim, Guinn, 2008; Roscoe, Bushunow, Jean-Pierre, et al., 2009). Although nondrug interventions may be helpful, the mainstay of PONV treatment is pharmacologic (Wilhelm, Dehoorne-Smith, Kale-Pradhan, 2007). The use of regional rather than general anesthesia is widely recommended to reduce PONV, although some surgical procedures (e.g., cesarean, some major orthopedic surgeries) are associated with a high incidence of PONV despite using regional anesthesia (Borgeat, Ekatodramis, Schenker, 2003).

• Administer PONV prophylaxis using one to two interventions to patients at moderate risk for PONV. For example, administer dexamethasone (Decadron) before anesthesia induction and a serotonin receptor antagonist (e.g., ondansetron [Zofran]) at the end of surgery. It is important to consider that research has shown that single-drug prophylaxis has a high failure rate and is associated with a resultant increased cost of care (Gan, Meyer, Apfel, et al., 2007; Watcha, 2000), so combinations of two antiemetics rather than a single antiemetic prophylactically are preferred.

• Administer PONV prophylaxis using two or more interventions (multimodal approach) to patients at high risk for PONV. For example, administer dexamethasone before anesthesia induction, IV total anesthesia (IVTA) propofol (Diprivan), a serotonin receptor antagonist at the end of surgery, and IV propofol rescue doses in the PACU. A prospective study of 376 patients at high risk for PONV revealed that the administration of three or more prophylactic antiemetics produced the largest reduction in PONV, but despite this aggressive treatment, 30% still experienced symptoms severe enough to interfere with function (White, O’Hara, Roberson, et al., 2008). A prospective observational study concluded similarly that compared with lower Apfel risk scores, a high Apfel risk score (see above) was associated with a higher incidence of emetic sequelae in the first 24 hours after surgery despite the prophylactic administration of multiple antiemetics (White, Sacan, Nuangchamnong, et al., 2008).

• Do not administer prophylactic antiemetic treatment to low-risk patients as this is not supported by current practice (Apfel, Korttila, Abdalla, et al., 2004; Gan, Meyer, Apfel, et al., 2003, 2007); however, provide antiemetic treatment to patients with PONV who did not receive prophylaxis or in whom prophylaxis failed.

Antiemetics

There are numerous antiemetic drug options available (Carlisle, Stevenson, 2006), and selection should be based on evidence of efficacy and safety as well as consideration of cost (see Table 19-1). A systematic review concluded that no one antiemetic agent is superior to another (Wilhelm, Dehoorne-Smith, Kale-Pradhan, 2007). A multicenter study of over 4000 patients at high risk for PONV (greater than 40% risk per Apfel risk scoring system [Apfel, Laara, Koivuranta, et al., 1999]) and undergoing a variety of types of surgeries were randomized to receive one of the following interventions: ondansetron (4 mg IV) or no ondansetron; dexamethasone (4 mg IV) or no dexamethasone; droperidol (Inapsine) (1.25 mg IV) or no droperidol; propofol or a volatile anesthetic (i.e., isoflurane, desflurane, or sevoflurane); nitrogen or nitrous oxide; and remifentanil or fentanyl. Because propofol has been shown to reduce PONV (see the paragraphs that follow), twice as many patients were assigned to the propofol group to ensure adequate power to compare treatments. All of the antiemetics were similarly effective; ondansetron, dexamethasone, and droperidol reduced risk by approximately 26%, propofol by 19%, and nitrogen by 12%. Droperidol, which has a “Black Box” warning for potential QTc prolongation and torsades de pointes, was safe (see later in the chapter for more on droperidol). The researchers pointed out that the clinical implication of their findings is that, because the interventions were similarly effective, the safest and least expensive treatment should be used first.

The glucocorticoid dexamethasone is an excellent choice antiemetic because numerous studies have shown it to be effective, safe, and inexpensive (Gan, Meyer, Apfel, et al., 2007) (see Table 19-1). It may be given prophylactically before induction of anesthesia as well as for established PONV (Golembiewski, Chernin, Chopra, 2005) and has been shown to have similar efficacy to the serotonin antagonist tropisetron, which is not available in the United States (Wang, Ho, Uen, et al., 2002). Dexamethasone is administered as a single 4 mg to 8 mg IV bolus dose most often and has been combined in multimodal treatment plans with a number of other agents including ondansetron (Zofran) (Paech, Rucklidge, Lain, et al., 2007; Pan, Lee, Harris, 2008), granisetron (Kytril) (Fujii, Saitoh, Tanaka, et al., 1999; Gan, Coop, Philip, et al., 2005), dolasetron (Anzemet) (Coloma, White, Markowitz, et al., 2002; Rusch, Arndt, Martin, et al., 2007), droperidol (Sanchez-Ledesma, Lopez-Olaondo, Pueyo, et al., 2002), metoclopramide (Reglan) (Wallenborn, Gelbrich, Bulst, et al., 2006), and haloperidol (Haldol) (Chu, Shieh, Tzeng, et al., 2008; Rusch, Arndt, Martin, et al., 2007). Adverse effects are rare with short-term use (Gan, Meyer, Apfel, et al., 2007).

The serotonin receptor antagonists dolasetron, granisetron, and ondansetron are most effective when administered at the end of surgery (Gan, Meyer, Apfel, et al., 2007) (see Table 19-1). The newest serotonin antagonist palonosetron (Aloxi) is approved for prevention of chemotherapy-induced nausea and has been shown to significantly decrease PONV in a dose-related manner (0.075 mg IV significantly better than 0.025 mg IV) when administered immediately prior to anesthesia induction (Candiotti, Kovac, Melson, et al., 2008). The serotonin receptor antagonists are often combined with other antiemetics in multimodal PONV prophylaxis regimens. This practice is supported by a meta- analysis of 33 randomized controlled trials (3447 patients), which concluded that various combinations of a serotonin antagonist with dexamethasone or droperidol were equally effective with similar adverse effects, and the combinations were more effective than a serotonin antagonist alone (Habib, El-Moalem, Gan, 2004). A 2001 systematic review of the literature found that the serotonin antagonists available at the time were similarly effective for treatment of established PONV and that lower doses were as effective as higher doses, so the lowest dose in the dosing range was recommended (Kazemi-Kjellberg, Henzi, Tramer, 2001).

The serotonin antagonists are reported to prevent postoperative vomiting better than nausea (Gan, Meyer, Apfel, et al., 2007; Kazemi-Kjellberg, Henzi, Tramer, 2001); however, an analysis of data from 5161 patients concluded that ondansetron prevents both symptoms equally well (Jokela, Cakmakkaya, Danzeisen, et al., 2009). Ondansetron 8 mg twice daily is effective in an orally disintegrating tablet formulation (Zofran ODT) (Gan, Franiak, Reeves, 2002; Grover, Mathew, Hegde, 2009; Hartsell, Long, Kirsch, 2005). The most common adverse effect of the serotonin antagonists is headache (Kazemi-Kjellberg, Henzi, Tramer, 2001).

The anticholinergic scopolamine delivered via a transdermal patch (Transderm-Scop, Transderm-V) has been shown to prevent PONV with minimal adverse effects when applied preoperatively (White, Tang, Song, et al., 2007) (see Table 19-1). One patch (1.5 mg) should be applied behind the ear preoperatively, taking into account its 2- to 4-hour onset of action, and it can provide relief for up to 72 hours. A second patch may be applied after the first is removed at 72 hours. Transdermal scopolamine may be combined with other antiemetics in a multimodal treatment plan (Kranke, Morin, Roewer, et al., 2002). A randomized study of 126 patients undergoing cosmetic surgery found that the transdermal scopolamine patch combined with IV ondansetron (4 mg) was more effective in reducing PONV than ondansetron plus a placebo patch (Sah, Ramesh, Kaul, et al., 2009). Another study also found the combination to be more effective than ondansetron alone (Gan, Sinha, Kovac, et al., 2009). Transdermal scopolamine has also been used to reduce nausea and vomiting associated with intrathecal morphine post–cesarean section (Harnett, O’Rourke, Walsh, et al., 2007). Adverse effects include dry mouth, sedation, and visual disturbances; older patients may be more sensitive to CNS adverse effects, such as dizziness and agitation (Golembiewski, Chernin, Chopra, 2005).

Research in the 1990s demonstrated that the IV sedative hypnotic propofol at subhypnotic doses (5 to 10 mg IV push q 4 to 6 h or 0.5 to 1 per hour continuous infusion) reduced the overall incidence of PONV in patients at high risk for PONV without untoward sedative or cardiovascular (CV) effects compared with placebo (Ewalenko, Janny, Dejonckheere, et al., 1996). Since then, the drug has gained in popularity as a component of multimodal approaches designed to reduce PONV (Eberhart, Mauch, Morin, et al., 2002; Scudieri, James, Harris, et al., 2000) (see Table 19-1). A randomized controlled study administered 90 patients undergoing laparoscopic cholecystectomy one of the following regimens: (1) a multimodal management strategy that involved the use of TIVA propofol plus droperidol and ondansetron, (2) IV propofol at induction followed by inspired isoflurane/nitrous oxide-based anesthesia, droperidol, and ondansetron, or (3) TIVA with no other antiemetic prophylaxis (Habib, White, Eubanks, et al., 2004). The droperidol (0.625 mg) was administered at induction and ondansetron (4 mg) was administered at the end of surgery in groups 1 and 2. Complete response rate (no PONV and no rescue antiemetic) at 2 hours and 24 hours after surgery was 90% and 80% in group 1, 63% and 63% in group 2, and 66% and 43% in group 3. Patient satisfaction was also higher in group 1. The researchers noted, however, that the higher cost of propofol compared with volatile anesthetics and its short-lived antiemetic effect (limited to early postoperative period) makes it suitable for use only in patients at very high risk for PONV.

Novel Approaches to the Management of PONV

Some novel and relatively inexpensive approaches have been tried for the management of PONV. Researchers applied nicotine patches to patients undergoing laparoscopic cholecystectomy under general anesthesia based on research that shows cigarette smoking reduces risk of PONV (Ionescu, Badescu, Acalovschi, 2007). Patients in this study (N = 75) were randomized according to their cigarette smoking status: (1) nonsmokers, (2) patients who had stopped smoking at least 5 years prior and received one 16.6 mg nicotine patch, and (3) patients who currently smoked. There was a 20% reduction in PONV in patients in group 2 compared with group 1 and no difference between group 2 and group 3.

Intraoperative supplemental oxygen has been suggested as a strategy to reduce PONV, but research is conflicting. One early study showed that the incidence of PONV was significantly reduced in patients following laparoscopic gynecologic surgery who received 80% supplemental intraoperative oxygen (22%) and those who received 8 mg of ondansetron at induction (30%) compared with patients who received 30% supplemental oxygen intraoperatively (44%) (Goll, Akca, Grief, et al., 2001). However, a later study showed that neither 30% nor 80% intraoperative oxygen administration reduced PONV in 100 patients undergoing ambulatory gynecologic laparoscopy (Purhonen, Turunen, Ruohoaho, et al., 2003). Intraoperative supplemental oxygen did not decrease the incidence of PONV post–cesarean section delivery with neuraxial anesthesia either (Phillips, Broussard, Sumrall, et al., 2007).

Studies have shown that wrist acustimulation/acupressure (acupuncture point P6 [pericardium 6]) (ReliefBand®) reduces PONV (Gan, Jiao, Zenn, et al., 2004; Lee, Fan, 2009; Nunley, Wakim, Guinn, 2008; Roscoe, Bushunow, Jean-Pierre, et al., 2009) and that when combined with ondansetron (4 mg), the response rate to acustimulation is increased and quality of recovery and patient satisfaction are improved (Coloma, White, Ogunnaike, et al., 2002; White, Issioui, Hu, et al., 2002). The optimal time to administer acustimulation for antiemetic prophylaxis is after surgery (White, Hamza, Recart, et al., 2005). Figure 19-2 shows the location of acupuncture point P6.

A meta-analysis of research on gum chewing during the early postoperative period following colectomy concluded that the practice may enhance GI recovery and reduce length of stay (Purkayastha, Tilney, Darzi, et al., 2008). Improved GI recovery may contribute to a lower incidence of PONV.

Effective, Safe, and Inexpensive Treatment

By far, the most effective, safest, and least expensive way to treat PONV is to reduce the opioid dose whenever possible. Postoperative opioid orders should include the expectation that nurses will consider decreasing the opioid dose by 25% prior to or in conjunction with pharmacologic treatment of moderate-to-severe PONV (see following patient example and Form 17-1 on p. 464 for an example of how decreases in opioid dose can be included in opioid order sets). Decreasing the opioid dose is facilitated by adding or increasing a nonopioid, such as an NSAID or acetaminophen, or adding a local anesthetic to the epidural opioid solution to provide additional pain relief. If patients are too nauseated to take oral nonopioids, they may be given rectally (see Section III and Chapter 14 in this section for more on rectal administration).

Approaches with Little or No Effectiveness

In their 2007 guidelines, the American Society for Ambulatory Anesthesia concluded that there is a lack or limited evidence of a prophylactic effect for metoclopramide (10 mg IV), ginger root, and cannabinoids for PONV (Gan, Meyer, Apfel, et al., 2007). Metoclopramide is reported to be no more effective than placebo for its treatment (Gan, 2002); however, one randomized controlled study (N = 3140) found doses of 25 to 50 mg of metoclopramide in combination with dexamethasone administered intraoperatively reduced the frequency of PONV without a high incidence of adverse effects (Wallenborn, Gelbrich, Bulst, et al., 2006). Doses this high are not recommended; the drug has been reported to produce extrapyramidal symptoms, even in low doses (i.e., 10 mg) (Moss, Hansen, 2008).

Although promethazine (Phenergan) has been used for many years and has efficacy for the treatment of established PONV (Habib, Breen, Gan, 2005; Gan, Meyer, Apfel, et al., 2007; Habib, Reuveni, Taguchi, et al., 2007; Moser, Caldwell, Rhule, 2006), it is associated with significant adverse effects, including excessive sedation, respiratory depression, dysphoria, dystonia, and extrapyramidal symptoms (McGee, Alexander, 1979; Sheth, Verrico, Skledar, et al., 2005). Further, significant tissue damage can occur when promethazine is given intravenously (Institute for Safe Medication Practices, 2006a, 2006b). A “Black Box” warning added to promethazine prescribing information describes these injuries and the risk of unintentional intra-arterial injection and recommends deep IM injection of the drug; SC injection is contraindicated (U.S. FDA, 2009c). It is common practice to administer promethazine based on the belief that it will enhance opioid analgesia; however, early research dispelled this misconception (McGee, Alexander, 1979), and the practice is discouraged (Pasero, Portenoy, McCaffery, 1999). If promethazine is used, low doses are recommended, particularly in older adults (Habib, Breen, Gan, 2005; Moser, Caldwell, Rhule, 2006). Doses of 6.25 mg were found to be as effective as higher doses (e.g., 12 mg) (Habib, Reuveni, Taguchi, et al., 2007).

Prochlorperazine (Compazine) provided better relief of nausea and vomiting than promethazine in patients in the emergency department (ED) (Ernst, Weiss, Park, et al., 2000). Prochlorperazine has a faster onset and causes less sedation than promethazine as well (Golembiewski, Chernin, Chopra, 2005).

Another commonly used drug is hydroxyzine (Vistaril), but the doses that would be required to produce analgesia create significant risk of respiratory depression that is not reversible by naloxone (Gordon, 1995). IM hydroxyzine is especially irritating to the muscle and soft tissue and can produce sterile abscesses, so this practice is discouraged as well. Other older drugs—dimenhydrinate (Dramamine) (Kothari, Boyd, Bottcher, et al., 2000) and diphenhydramine (Benadryl)—are occasionally used for PONV but can cause significant sedation and dizziness. Antiemetics with better efficacy and safety should be considered before these drugs are used for treatment of PONV (Gan, Meyer, Apfel, et al., 2007) (see Table 19-1).

The routine use of NG tubes during surgery as a means of reducing PONV is not recommended. A large case control study (N = 4055) evaluated the association between NG tube use and the incidence of nausea, emesis, and overall PONV and showed no reduction in the incidence of these three outcome measures; the incidence of PONV was 44.4% with intraoperative NG tube versus 41.5% in controls (Kerger, Mascha, Steinbrecher, et al., 2009).

The Meperidine Misconception

Although institutional quality improvement initiatives have resulted in a significant decline in the use of meperidine over the years (Gordon, Jones, Goshman, et al., 2000), the drug continues to be a first-choice analgesic of many prescribers (Seifert, Kennedy, 2004). This is particularly true for the management of pain during GI procedures and in patients with pancreatitis or cholecystitis; however, there are many disadvantages to the use of meperidine for pain management, including accumulation of its toxic metabolite with repeated dosing and its inappropriateness in older adults (Latta, Ginsberg, Barkin, 2002) (see Chapter 13).

A review of the literature concluded that morphine may be of more benefit than meperidine by offering analgesia without the risks associated with meperidine and that there are no studies or evidence to indicate morphine is contraindicated for use in acute pancreatitis (2001). Low-dose transdermal fentanyl (12.5 to 25 mcg) caused no significant changes in sphincter of Oddi pressure in patients with pancreatitis in one study and was described as the ideal analgesic for treatment of pancreatitis pain (Koo, Moon, Choi, et al., 2009). An interesting letter to the editor suggested that the choice of meperidine over other opioids is not based on evidence and persists because of the perpetuation of the misconception that meperidine has no effect on the sphincter of Oddi (Lee, Cundiff, 1998). The authors equated this to the “medical equivalent of an urban legend.” The best course of action is to avoid meperidine for all types of pain, including procedural pain and pancreatitis pain, and rely instead on other mu agonist opioids, such as morphine, hydromorphone, or fentanyl.

Assessment of Pruritus

The use of a self-report tool to assess pruritus may be helpful, particularly in cases of intractable pruritus. A 4-point verbal rating scale (VRS-4) and an 11-point verbal numeric rating scale (VNRS-11) are used most often (Jenkins, Spencer, Weissgerber, 2009). The VRS-4 matches a score to the patient’s description of itch intensity: 0 = no itching; 1 = mild itching; 2 = moderate itching; 3 = severe itching; the VNRS-11 is used similarly to the numerical rating scale (NRS) for pain intensity assessment: 0 = no itch and 10 = the worst imaginable level of itching. A study of 50 parturients demonstrated a strong correlation between the two scales, leading the researchers to conclude that each verbal descriptor on the VRS-4 could be substituted with a quantifiable range on the VNRS-11, i.e., 1 to 3 = mild itching; 4 to 7 = moderate itching; and 8 to 10 = severe itching (Jenkins, Spencer, Weissgerber, 2009).

Pharmacologic Management of Pruritus

Antihistamines such as diphenhydramine relieve only itch caused by histamine release, such as insect bites, urticaria, and allergic skin reactions. Although they are widely used, there is no strong evidence that antihistamines relieve opioid-induced pruritus (Grape, Shug, 2008). Some suggest that they may be more effective for pruritus caused by systemic opioids than neuraxial opioids, but research is lacking (Waxler, Dadabhoy, Stojiljkovic, et al., 2005). Patients may report being less bothered by itching after taking an antihistamine, but this is likely the result of sedating effects (Ho, Gan, 2009). Sedation can be problematic in those already at risk for excessive sedation, such as postoperative patients, as this can lead to life-threatening respiratory depression (Anwari, Iqbal, 2003) (see later in this chapter). Thus careful monitoring of sedation levels is recommended when antihistamines are combined with opioid administration, and they should not be administered if patients are excessively sedated.

Prostaglandins do not elicit pruritus when applied to the skin, but they do act synergistically with histamine to potentiate the histamine-elicited itch (Waxler, Dadabhoy, Stojiljkovic, et al., 2005). Topical ketorolac tromethamine has been shown to relieve ocular itching, most likely through its inhibition of prostaglandin synthesis (Donshik, Pearlman, Pinnas, et al., 2000) (see Section III).

It is thought that serotonin (5-HT) acts on 5-HT₃ receptors to generate the sensation of pruritus (Waxler, Dadabhoy, Stojiljkovic, et al., 2005). This helps to explain why the serotonin receptor antagonists have been used to successfully treat pruritus caused by intraspinal opioids. A prospective study randomized 105 patients to receive ondansetron 4 mg, dolasetron 12.5 mg, or placebo 30 minutes before spinal morphine and bupivacaine anesthesia (Iatrou, Dragoumanis, Vogiatzaki, et al., 2005). Patients in both treatment groups experienced significantly less pruritus during the first 8 postoperative hours than the placebo group, and severe pruritus was noted only in those who received placebo. Others have reported similar results with ondansetron for prevention of pruritus from intrathecal morphine (Yeh, Chen, Lin, et al., 2000) and fentanyl (Gurkan, Toker, 2002). However, no benefit was found with the combination of dexamethasone (8 mg) and ondansetron (4 mg) for prevention of intrathecal morphine-induced pruritus (Szarvas, Chellapuri, Harmon, et al., 2003). Ondansetron orally disintegrating tablets (ODT) 8 mg, ondansetron IV 4 mg, or placebo was administered prior to intrathecal morphine in 150 men undergoing surgery, and although there were no significant differences in PONV, the incidences of pruritus were 56%, 66%, and 86% in the ODT, ondansetron IV, and placebo groups, respectively (Pirat, Tuncay, Torgay, et al., 2005).

Established pruritus is responsive to treatment with ondansetron as well. A randomized controlled study of 80 women with moderate to severe pruritus following intrathecal morphine for cesarean section were given ondansetron (4 mg) or placebo (Charuluxananan, Somboonviboon, Kyokong, et al., 2000). Relief was achieved in 80% of those who received ondansetron compared with 36% of those who received placebo. The recurrence rates within 4 hours after administration were 12% and 70% for ondansetron and placebo, respectively. Nausea and vomiting were also significantly less in the ondansetron group. A case report described the treatment of intractable pruritus with 4 mg of IV ondansetron following spinal fentanyl and bupivacaine anesthesia (Henry, Tetzlaff, Steckner, 2002).

An interesting study found that gabapentin (Neurontin) was effective in reducing the incidence of intrathecal morphine-induced pruritus (Sheen, Ho, Lee, et al., 2008). Patients (N = 86) were randomized prior to limb surgery to receive 1200 mg of gabapentin or placebo orally 2 hours prior to surgery. The incidence of pruritus was 47.5% in the gabapentin group and 77% in the placebo group. The severity of pruritus was greater, its onset shorter, and duration longer in the placebo group. The researchers suggested that the effectiveness of gabapentin may be related to its action on central neurons and the fact that intrathecal morphine-induced pruritus is a “neurogenic itch.” Anticonvulsants such as gabapentin are used to treat pruritus associated with advanced illness as well (Pittelkow, Loprinzi, 2004).

Subanesthetic doses of IV propofol (e.g., 10 mg) have been shown to relieve pruritus associated with intrathecal morphine and are suggested as a second-line option after administration of a serotonin receptor antagonist (Grape, Schug, 2008). One study randomized 50 postoperative patients who had intrathecal morphine-induced pruritus to receive 10 mg of IV propofol or placebo; a dose was repeated 5 minutes later in patients who had a pruritus score of more than 2 on scale of 0 to 5 (Borgeat, Wilder-Smith, Saiah, et al., 1992). Treatment failure was defined as the persistence of a pruritus score of greater than 2 at 5 minutes after treatment. The success rate was 86% and 16% in the propofol and placebo groups, respectively. In contrast, a later randomized controlled study of 29 women with intrathecal morphine-induced pruritus following cesarean section found no difference between propofol 10 mg IV and placebo (Beilin, Bernstein, Zucker-Pinchoff, et al., 1998).

Topical local anesthetics such as EMLA or lidocaine patch 5% may provide relief of pruritus in some patients. These preparations are not convenient for generalized pruritus, but may be helpful for localized areas that are particularly bothersome (Pittelkow, Loprinzi, 2004).

Switching to Another Opioid

Switching to another opioid may relieve pruritus but is usually reserved for patients with pruritus that has been unresponsive to other treatments. A case report described intractable itching that began when morphine was initiated in a patient with small round cell tumor of the pelvis and was not improved when the patient was switched to fentanyl and then to hydromorphone (Tarcatu, Tamasdan, Moryl, et al., 2007). Diphenhydramine and hydroxyzine also had no effect. When oxycodone and a low-dose naloxone infusion (0.25 mcg/kg/h) were started, the patient’s pruritus dramatically improved. Pain control was achieved with titrated doses of oxycodone (60 mg every 3 hours) and IV hydromorphone PRN. Naloxone was discontinued seven days later with no recurrence of itching. The authors suggested that the resolution of pruritus was because the oxycodone created a balance between mu and kappa opioid receptors through a predominantly kappa agonist effect. (Morphine, hydromorphone, and fentanyl bind primarily to mu opioid receptors.) Further, the authors discounted a singular role for naloxone since itching did not recur after it was discontinued.

Effective, Safe, and Inexpensive Treatment

A common clinical observation is that patients with postoperative opioid-induced pruritus have well-controlled pain. This may be because, as mentioned, the two sensations of pain and itch seem to be interrelated; painful stimuli can inhibit itching, and inhibition of pain processing may enhance itching. This helps to explain why a small decrease in the opioid dose is such an efficient solution to opioid-induced pruritus. Opioid dose reduction is by far the single most effective, safest, and least expensive treatment for pruritus. Postoperative opioid orders should include the expectation that nurses will decrease the opioid dose by 25% prior to or in conjunction with pharmacologic treatment of moderate-to-severe pruritus (see Form 17-1 on p. 464 for an example of how decreases in opioid dose can be included in opioid order sets). Patients usually tolerate this small reduction in opioid dose without any loss of analgesia and experience a significant reduction or complete resolution of their pruritus. Decreasing the opioid dose is facilitated by adding or increasing a nonopioid, such as an NSAID or acetaminophen, or adding a local anesthetic to the epidural opioid solution to provide additional pain relief.

In summary, pruritus should be treated according to its severity or prevented based on its expected severity (see Table 19-1). The use of a self-report pruritus assessment scale is recommended to help determine severity and effectiveness of treatment in patients with intractable pruritus. Intraspinal opioid-induced pruritus should be prevented with administration of a serotonin antagonist, particularly when a single-bolus intraspinal technique is used. Serotonin antagonists may be helpful for established pruritus as well. Mild facial and chest pruritus may be relieved by cold compresses. The easiest and most effective treatment for pruritus is to decrease the opioid dose if possible. Opioid agonist-antagonists and opioid antagonists should be reserved for severe pruritus, and the patient should be watched closely for any increase in pain if these are used.

Postoperative Delirium

The reported incidence of postoperative delirium is 10% to 60%, and it is most frequent in older adults (Vaurio, Sands, Wang, et al., 2006). This number is likely to increase as more old than young adults have surgery (Amador, Goodwin, 2005). Delirium is associated with prolonged hospital stay, greater use of hospital resources, and increased cost of care (Demeure, Fain, 2006). Despite this, there is little well-controlled research on the incidence and prevention of delirium in the hospitalized patient (Siddiqi, Holt, Britton, et al., 2007).

Similar to the terminally ill, delirium tends to be underdiagnosed in the postoperative setting (Amador, Goodwin, 2005). It is characterized by alterations in orientation, consciousness, memory, thought processes, and behavior (Demeure, Fain, 2006). Patients may demonstrate either hyperactivity or hypoactivity and may be lucid at times, but typically the symptoms are worse at night (“sundowning”). A variety of delirium screening and diagnostic tools are available (see previous discussion and Form 19-1). Table 19-1 provides treatment suggestions.

Opioids are often blamed for the occurrence of postoperative confusion; however, little research supports the assertion that they are a major cause. In fact, pain and other factors, such as sleep disturbance, type of surgery, choice of opioid, and method of administering opioids, have been directly linked to postoperative cognitive impairment. As described in the terminally ill, the importance of an accurate assessment with consideration of the patient’s baseline cognitive status is critical in the confused postoperative patient; many patients undergoing surgery have pre-existing confusion and cognitive impairment and should be assessed as “at risk” for postoperative confusion. Box 19-2 lists the risks associated with postoperative delirium.

Poorly managed pain appears to be an important risk factor for postoperative delirium. Early research showed that a decline in mental status was linked to poor pain control in patients aged 50 to 80 in one study (Duggleby, Lander, 1994). Pain, not analgesic intake, predicted mental decline. More recently, a study was conducted in 330 consecutively admitted patients (age 65 years or older) scheduled for cardiac surgery (Vaurio, Sands, Wang, et al., 2006). Interviews and baseline assessments of neurologic, functional, and cognitive status and pain were completed within 48 hours before surgery and repeated at 24 and 48 hours after surgery. Both the presence of postoperative pain and increased pain postoperatively were found to be independent predictors of postoperative delirium in these patients and to have a greater impact on the development of delirium than any of the other factors evaluated (e.g., age, type of anesthesia, postoperative medications, and preoperative cognitive status). There was also an ordered relationship between levels of postoperative pain and risk of delirium; those with severe pain were at a greater risk of developing delirium than those with moderate pain. An important finding was that neuraxial analgesia and IV PCA carried the same risk despite the lower opioid doses given neuraxially. Similarly, an earlier study reported no difference in the incidence of delirium between IV fentanyl and epidural fentanyl analgesia for pain following total knee replacement (Williams-Russo, Urquhart, Sharrock, et al., 1992). In contrast, fewer complications, including mental confusion, were found in older adult men receiving IV PCA morphine compared with those receiving IM morphine (Egbert, Parks, Short, et al., 1990).

The choice of opioid used to control pain also may influence the incidence of mental confusion, but research demonstrating differences among the various opioids is lacking. One exception is meperidine, which is more likely than other opioid drugs to cause delirium in postoperative patients of all ages (Fong, Sands, Leung, 2006). In a case control study (N = 91 with 1 to 2 controls), meperidine more than doubled the risk of delirium when given either epidurally or IV (Marcantonio, Juarez, Goldman, et al., 1994). It has also been shown to be the only opioid to have a negative impact on mood (Latta, Ginsberg, Barkin, 2002), sometimes the first sign of neurotoxicity. Early research found that IV PCA fentanyl produced less depression of postoperative cognitive function than IV PCA morphine (Herrick, Ganapathy, Komar, et al., 1996). IV PCA morphine and IV PCA hydromorphone provided similar analgesic efficacy, adverse effects, mood, and cognitive function in another study (Rapp, Egan, Ross, et al., 1996). Although the patients receiving hydromorphone experienced improved mood, they also demonstrated poorer cognitive performance than those receiving morphine.

In summary, these research findings reinforce that factors other than dose of opioid, such as quality of pain control, play a more important role in the development of delirium. Studies and clinical observations suggest that improvement in pain management practices, including the administration of appropriate doses of opioids, is one way to reduce confusion after surgery. The common reaction of many clinicians to abruptly discontinue the opioid when a patient develops confusion may worsen the confusion and is not recommended. After careful evaluation of all potential causes, a better approach is to add or increase a nonopioid or nonsedating adjuvant analgesic and decrease the opioid dose if a decrease is thought to be necessary (see Table 19-1).

Opioids and Driving Ability

For many people, driving an automobile is a daily activity that is a key element to maintaining their independence, engaging in employment, and remaining socially active. They are essentially disabled if they are deprived of the freedom to drive. As the use of long-term opioid therapy gains greater favor, especially in the care of patients with persistent noncancer pain, it is of great importance to the patient, physician, and public to determine their ability to drive safely.

This problem has been studied since the early 1960s when methadone maintenance therapy was introduced (Dole, Nyswander, 1965). One of the first studies compared the driving records of 401 patients before and after methadone maintenance therapy with 182 drivers selected at random (Gordon, 1973). During methadone treatment, patients did not differ significantly from the comparison group in rates of accidents or in convictions for driving offenses. Interestingly, while using heroin the patients had had fewer accidents, a finding that had no satisfactory explanation. (Of interest, another similar study found that heroin users had a slightly better driving record on heroin than they did on methadone [Maddux, Williams, Ziegler, 1977]). Gordon’s conclusion was the same as many studies to follow: that taking an opioid (methadone in this case) on a long-term basis does not have a significant detrimental effect on the ability to drive.

Before embarking on an exploration of the effects on driving ability of long-term opioid therapy for persistent pain, it is important to point out that those patients who take opioids short-term, those in whom long-term opioid therapy is being initiated, and those on long-term therapy in whom opioid doses are being increased are different circumstances. Studies of the psychomotor and cognitive effects of opioids in these groups are generally different from those discussed in relation to long-term stable opioid therapy.

Extensive reviews of the literature have examined studies on the effects of a single opioid dose or an increased dose of opioid in patients on long-term therapy (Chapman, Byas-Smith, Reed, 2002; Zacny, 1996). These reviews reveal that this is a complex topic and that results of various studies are inconclusive. Results depend upon a multitude of factors such as type of opioid administered, dose of opioid, use of other medications, and whether or not the patient is in pain. For example, it appears that in opioid-naïve, healthy volunteers without pain, psychomotor and cognitive functioning are impaired more by some opioids (e.g., mixed-agonist-antagonists such as buprenorphine and pentazocine) than others (e.g., codeine and morphine). However, this may have little relevance to patients with pain, since pain itself impairs cognition (Lorenz, Beck, Bromm, 1997), and pain relief may improve cognitive functioning (Jamison, Schein, Vallow, et al., 2003). For example, treatment of neuropathic pain with both low and high doses of levorphanol resulted in improvement in some measures of cognitive performance (Rowbotham, Twilling, Davies, et al., 2003).

In a landmark study of patients with cancer, the cognitive effects of stable doses of opioids compared with changes in opioid dose in patients on long-term opioid therapy was conducted by Bruera and others (Bruera, Macmillan, Hanson, et al., 1989). Twenty inpatients with cancer who had increased their opioid dose by at least 30% in the 3 days before testing were compared with 20 other inpatients who had been on a stable opioid dose for at least 7 days. A variety of opioids were used. Compared with the group on a stable opioid dose, the group taking the recently increased dose reported more drowsiness after the dose, and significantly more cognitive impairment occurred 45 minutes after opioid administration compared with several hours afterwards.

The immediate cognitive effects following a dose of opioid are often observed by clinicians and are supported by other studies. The immediate effects of 130 mg of dextropropoxyphene showed impairment was greater when measured 2 hours after than 4 hours after a dose (Saarialho-Kere, Julkunen, Mattila, et al., 1988). Thinking that the immediate effects of an opioid will persist seems to cause some clinicians to fear long-term use of opioids. In correcting this misconception, an explanation of the effects of tolerance may be helpful.

Cognitive performance was examined in a study comparing 6 female patients with persistent noncancer pain before and after being treated with modified-release morphine (Lorenz, Beck, Bromm, 1997). The first testing was before initiation of morphine, and the second examination was after the patients reported sufficient pain relief and had been on a stable dose for at least 3 days. Doses ranged from 30 to 150 mg of modified-release morphine per day. During the second testing session, behavioral and physiologic indicators of vigilance and cognitive function revealed lack of sedation and improved alertness. The authors suggested that these effects might be due in part to the possibility that pain relief made it unnecessary to constantly shift attention between pain and other activities.

The above studies and others have resulted in the recommendation that after a person’s stable opioid dose is increased, caution should be exercised during the first few hours after taking the dose and until the patient has been on a stable dose for 7 days. Cognitive performance during the first few days of opioid use and during the first few hours after a given dose seems likely to be worse. However, after about 7 days of stable opioid dosing, cognitive effects such as sedation seem to subside.

Effects of long-term opioid therapy on driving ability were studied by comparing 16 patients with persistent noncancer pain on opioids for 6 months or longer with 327 cerebrally compromised patients from another study who had off-road and on-road testing for driving ability (Galski, Williams, Ehle, 2000). In the study of the cerebrally impaired patients, certain off-road tests were found to predict success or failure in the on-road evaluation. The patients on opioids were given the off-road tests, and the results were compared with those of the cerebrally impaired patients who had undergone the same evaluation and then passed or failed the on-road evaluation. Generally the patients on opioid therapy performed better than the cerebrally compromised patients who passed the on-road test. Patients on long-term opioid therapy had faster reaction times and did not manifest any major problems with coordination, but they had greater difficulty following instructions and a tendency toward impulsivity, that is, hurrying and sacrificing accuracy for speed. Nevertheless, the results give general support for the ability of patients on long-term opioid therapy to pass actual on-road driving tests.

Twenty-nine patients with persistent noncancer pain (primarily low back pain) on long-term modified-release oxycodone were compared with 90 healthy volunteers to assess their driving ability (Gaertner, Radbruch, Giesecke, et al., 2006). Attention reaction, visual orientation, motor coordination, and vigilance were evaluated. This included a test battery that followed the German national recommendations on tests to determine driving ability. Permission to drive is usually denied if the score is below the sixteenth percentile. Using this definition of driving ability, there was no significant difference between the patients receiving modified-release oxycodone and the control group. The authors concluded that long-term treatment with modified-release oxycodone did not prohibit driving. However, they recommended individual assessment.

Another study using the German national recommendations on tests to determine driving ability compared 90 healthy controls with 21 patients with noncancer pain who had been treated with transdermal fentanyl for at least 4 weeks without a dose change in the last 12 days (Sabatowski, Schwalen, Rettig, et al., 2003). The median dose of transdermal fentanyl was 50 mcg/h with a range of 25 to 400 mcg/h. Of the 21 patients on transdermal fentanyl, all but 1 considered themselves fit to drive while the one “didn’t know.” Based on test results, the authors concluded that patients with noncancer pain on a stable dose of transdermal fentanyl do not have any clinically significant impairment of psychomotor or cognitive function that would prevent them from driving a car.

Of note is the fact that in the above study all patients except 1 correctly assumed they were able to drive. This suggests the wisdom of asking patients about their ability to drive and also of comparing these answers in future studies to actual ability to drive.

Driving performance, cognition, and balance were also studied before and after adding transdermal fentanyl to treatment regimens for 23 patients with persistent noncancer pain (Menefee, Frank, Crerand, et al., 2004). After patients were stabilized on a dose for 1 month, the tests were repeated. The median dose was 50 mcg/h. Once the dose of transdermal fentanyl had stabilized, it did not have a negative effect on driving performance, reaction time, or cognition. In fact, there was improvement on some measures of cognitive performance.

A poster presentation (Furlan, Lakha, Yegneswaran, et al., 2007) reported on the results of a literature search for controlled studies of long-term opioid use and the effect on driving. Although 1720 studies were located, only 31 were usable. They found that only one study tested driving on the road and two on a driving simulator and that, in these studies, stable opioid doses did not impair driving ability. In all 31 studies, there were 79 evaluations, and 27 found impairment, but no details about this impairment were given. The authors concluded that the quality of the studies was very low, but stated, “stable doses of opioids do not affect driving performance in a chronic pain population” (p. 127).

A structured evidence-based review, which is different from a meta-analysis, was conducted to examine whether long-term stable opioid dosing caused impairment in driving ability (Fishbain, Cutler, Rosomoff, et al., 2003). A literature search revealed 209 references, but only 48 were relevant to addressing driving abilities of patients on stable doses of opioids. These articles were reviewed in detail and then sorted and tabulated in several topic areas such as psychomotor abilities, cognitive functioning, and motor vehicle violations. Based on guidelines developed by the Agency for Health Care Policy and Research (AHCPR), the studies were categorized according to various levels of evidence, from well-designed controlled studies to case reports, and then categorized as to their strength and consistency of evidence. Following are the results of this evidenced-based review of studies of patients on long-term stable doses of opioids:

The authors of the above study concluded that overall the majority of studies appeared to indicate that patients on stable doses of long-term opioid therapy are not impaired in regard to driving skills. The authors also pointed out that the most relevant evidence were those studies that showed that opioid-tolerant patients appear to perform driving skills as well as controls, as stated in number 5 above. Nevertheless, they recommended that the decision to drive be made on an individual basis and specified the recommendations in Box 19-3.

A structured evidence-based literature review using methodology similar to the above study was conducted to determine if there was evidence of an association between opioid use and intoxicated driving, motor vehicle accidents (MVA), and MVA fatalities (Fishbain, Cutler, Rosomoff, et al., 2002). Except for those subjects identified as being on methadone maintenance therapy, no information was provided on how long the subjects had been taking opioids. The evidence indicated that opioids probably are not related to intoxicated driving, MVA, or MVA fatalities, supporting the recommendation that patients on opioids be allowed to drive. However, again, this decision should be made on an individual basis.

Compared with patients with noncancer pain on long-term opioid therapy, determining if patients with cancer on long-term opioid therapy are able to drive is a more complex matter (Brandman, 2005). Patients with cancer often have co-morbidities, neurologic changes, other medications, and advancing disease, causing their driving ability to change over time. The reader is referred to the specific guidelines given in the Brandman (2005) article for helping patients and physicians determine safe driving ability in patients with cancer on long-term opioid therapy. Some of these suggestions include asking occupational therapists to perform a driving assessment of the patient and considering the use of a CNS stimulant such as methylphenidate or modafinil for patients who report drowsiness.

In a medicolegal case report that presented the details of the case of a patient with persistent low back pain on long-term opioid therapy, the authors submitted some recommendations on how pain physicians should approach the problem of whether patients should be allowed to drive while on long-term opioid therapy (Fishbain, Lewis, Cole, et al., 2007). Although their literature review indicated that such patients could drive safely once they had developed a tolerance to the sedating effects of opioids, they offered specific guidelines such as suggesting that some form of informed consent concerning the risks of driving while on long-term opioid therapy should be obtained. Specific points they felt should be discussed with the patient on long-term opioid therapy included advising the patient of the current literature on driving while on this therapy, that whether or not to drive was their own personal decision, and that if they decided to drive they should obey certain rules. Some of these rules were not to drive for 4 to 5 days after beginning opioid treatment, not to drive if they felt sedated, and not to use alcohol or other illicit drugs before driving. The reader is referred to this article for information on other recommendations. Similar recommendations are available in Fishbain, Cutler, Rosomoff, et al., 2003 (see Box 19-3).

Not only patients and their families but also policymakers should be informed that the majority of the current literature shows that driving ability is not necessarily impaired in patients on stable doses of opioids. Without this knowledge, those responsible for enacting laws related to driving may restrict patients taking opioids. For example, in 1995 a guide was distributed to the British medical community recommending that people taking morphine-like opioids should not drive (as reported in Zacny, 1996). Zacny (1996) pointed out that the science of opioids and functioning should be used when making policy decisions, rather than hearsay and speculation.

In summary, whether or not a patient taking opioid analgesics can drive must be decided on an individual basis. Some patients will be able to drive safely, but others may be impaired by opioids. Impairment may be brief such as at the beginning of opioid therapy or following significant increases in dose. Some opioids cause more impairment for an individual patient than others, and a change in opioid may remedy the situation. It seems prudent to warn all patients that driving ability may be compromised and that they should not drive if they feel drowsy, dizzy, or impaired in any way.

It should be noted that the concern over effects of medication on driving ability should not be limited to opioids. Currently there seems to be a bias against the use of opioids that has helped fuel the concern about how these analgesics affect driving ability. However, other medication deserves attention too. Studies of older adults indicate that medications such as antidepressants or benzodiazepines (Ray, Fought, Decker, 1992), especially long–half-life benzodiazepines (Hemmelgarn, Suissa, Huang, et al., 1997) are at greater risk for motor vehicle crashes than those on opioids. Cognitive impairment was studied by comparing 13 patients taking benzodiazepines only with 13 patients taking opioids only (Hendler, Cimini, Long, 1980). The results showed that benzodiazepines were far more likely to produce cognitive impairment. In a study of 40 licensed drivers, age 25 to 44, the effects of fexofenadine (60 mg, Allegra), diphenhydramine (50 mg, Benadryl), alcohol (approximately 0.1% blood alcohol concentration), and placebo on driving performance were studied (Weiler, Bloomfield, Woodworth, et al., 2000). Driving performance was the poorest after taking diphenhydramine, impairing driving even more than alcohol.

Sedation Assessment

The observation that excessive sedation precedes opioid-induced respiratory depression (Abou hammoud, Simon, Urien, et al., 2009; Taylor, Voytovich, Kozol, 2003) indicates that systematic sedation assessment is an essential aspect of the care of opioid-naïve patients receiving opioid therapy (Nisbet, Mooney-Cotter, 2009; Pasero, 2009b). The importance of monitoring sedation to prevent clinically significant respiratory depression cannot be overemphasized. As the American Pain Society (APS) (2003) succinctly states, “No patient has succumbed to (opioid-induced) respiratory depression while awake” (p. 35).

Nursing assessment of opioid-induced sedation is convenient, inexpensive, and takes minimal time to perform (Pasero, 2009b). To assess sedation, the nurse should ask the patient a simple question, such as “What did you have for breakfast today?” then observe the patient’s ability to stay awake and answer the question. Patients who are excessively sedated will typically have difficulty keeping their eyes open and will fall asleep while responding, usually in the middle of a sentence. Nursing technicians (and other health care team members, including ancillary staff such as pastoral care counselors and dieticians) should be trained to watch for and report these signs of excessive sedation if evident and to routinely observe how alert patients are every time they obtain vital signs. Touching the patient to provide care and even simply entering the room can arouse a patient, giving the false impression of an acceptable level of sedation. Therefore, it is essential that the nurse or nursing technician observe the patient without stimulation for long enough to ensure an accurate evaluation.

Coadministration of Other Sedating Drugs

The administration of IM opioids to opioid-naïve patients receiving IV or epidural analgesia is unnecessary, can result in excessive sedation and clinically significant respiratory depression due to unpredictable absorption (see Chapter 14), and should be avoided. Instead, patients who are receiving IV or epidural opioid analgesia and experiencing uncontrolled pain should be carefully reevaluated. Clinician-administered IV or epidural supplemental doses and increases in the dose of the mainstay opioid usually are indicated (see Chapters 15 and 17).

Although not necessarily contraindicated with parenteral and intraspinal analgesia, muscle relaxants and anxiolytics can produce sedation. Therefore, prior to administering them, it is best to obtain approval for their administration from the prescriber responsible for the pain management plan (e.g., anesthesia provider with postoperative epidural analgesia). Antihistamines are often used to treat pruritus but also must be used with caution as they can be extremely sedating (see earlier discussion and Section V). Sedation can be problematic in those already at risk for excessive sedation, such as postoperative patients, as this can lead to life-threatening respiratory depression (Anwari, Iqbal, 2003). When these drugs are administered, increased monitoring of sedation is recommended, and they should not be administered if patients are excessively sedated.

Patients with Severe Pain and Excessive Sedation

It is important to note that the presence of sedation does not necessarily mean that patients are comfortable. Research shows that caring for patients with both severe pain and excessive sedation is common and presents major challenges to providing safe opioid therapy. In a time-course study of sedation and analgesia in 73 patients in the PACU after major surgery, 52 (73%) patients became sedated during titration necessitating discontinuation of titration; 21 remained awake, and titration was continued (Paqueron, Lumbroso, Mergoni, et al., 2002). Among those in whom titration was stopped because of sedation, 25% had a pain level higher than 50 on the VAS and only 50% had satisfactory pain relief (VAS less than 30). The findings of this study are important to consider when titrating opioids, i.e., that sedation can occur before pain is completely relieved and that sleep during opioid titration is not normal sleep but primarily the result of the sedative effects of the opioid.

Another time course study was conducted in 228 patients in the PACU following major orthopedic surgery who were titrated to either pain relief (VAS of 3 or less) or excessive sedation (Ramsay score higher than 2—a score of 3 indicates response to verbal commands only) (Abou Hammoud, Simon, Urien, et al., 2009). Titration was stopped when analgesia was achieved or an adverse effect occurred such as respiratory rate less than 12 breaths/min, oxygen saturation less than 95%, or sleep (sedation). Titration was stopped in nearly one half (45%) of the patients because of excessive sedation, but no serious adverse events (e.g., respiratory depression) occurred. The effectiveness of morphine was found to be increased in patients who had a decreased delay between extubation and titration, low initial VAS, and intraoperative multimodal analgesia (NSAID and/or acetaminophen administration). Analgesia was attained with a mean of 4 bolus doses and a mean morphine dose of 11 mg. The researchers emphasized that a multimodal approach is essential and that patients may benefit from earlier initiation of nonopioids (e.g., preoperatively). They proposed that higher amounts of bolus doses (e.g., 4 to 5 mg) rather than smaller bolus doses given at shorter intervals may be more effective in controlling severe pain, but appropriately pointed out that this could also increase adverse effects and compromise safety. An important finding was that 20 mg of titrated morphine was associated with a high incidence of excessive sedation, underscoring the need to use other approaches to manage pain that will allow lower opioid doses (i.e., multimodal analgesia). The effectiveness of stopping titration or reducing opioid dose when sedation is excessive is also demonstrated in this study by the lack of clinically significant respiratory depression.

A case control study demonstrated similar challenges in managing patients who have severe pain and also are excessively sedated. Patients in the PACU (N = 285) were titrated with IV morphine (2 to 3 mg) boluses every 5 minutes and then evaluated at 24 hours postoperatively (Lentschener, Tostivint, White, et al., 2007). PACU titration was discontinued when pain relief (VAS less than 3) was achieved or if adverse effects occurred such as excessive sedation (Ramsay score more than 3, sluggish or no response to loud voice or pain). Of the 285 patients, 26 were discharged from the PACU excessively sedated (mean Ramsay score 4) with uncontrolled pain (mean VAS 6.5). These patients were matched with 52 patients who were discharged not excessively sedated with adequate pain relief. The sedated patients received significantly more morphine in the PACU (mean 12 mg vs. 5 mg), had a two-times longer stay in the PACU (mean 240 min vs. 120 min), more often recalled having moderate-to-severe pain in the PACU, and had poorer quality sleep and higher pain scores at 24 hours after surgery.

An observational study compared pain, oxygen saturation levels, and sedation levels in three groups of patients: those receiving conscious sedation for colonoscopy (N = 18), those receiving IV PCA for postoperative pain management (N = 15), and those receiving nurse-controlled analgesia (PRN schedule) (N = 20) (Taylor, Voytovich, Kozol, 2003). The researchers reported that the analgesics and sedatives given were chosen by surgeons or anesthesiologists and adjusted to optimize analgesia, sedation, and safety, but exact drugs and doses were not provided in the publication of this study. Patients in group 2 (PCA) had trends in sedation levels greater than or equal to patients in group 1 (conscious sedation) during the first 10 hours postoperatively, and 72.7% of those receiving PCA were excessively sedated within the first 4 hours after discharge from the PACU. The lowest level of sedation was in group 3 (nurse controlled). Of interest, oxygen saturation was well maintained in all groups throughout the study. Pain scores declined steadily during the study in patients in groups 1 and 2 (average mean maximal pain score = 6.5/10) and undulated in group 3 (average mean maximal pain score = 5.5).

These studies add to a growing body of research showing that sedation precedes analgesia and despite being sedated, some patients will report pain. Nevertheless, opioid doses should not be increased (titration should be stopped) in patients who are excessively sedated, and they should be watched closely until they are no longer sedated. The APS (2003) recommends close monitoring of sedation and respiratory status for at least 3 hours after the peak of the last dose administered. Other options to achieve comfort such as adding or increasing the dose of nonopioids should be implemented. See the following discussion and Table 19-1 for the recommended approach for prevention and treatment of sedation and respiratory depression in these complex patients.

Treatment of Excessive Sedation

Opioid-induced sedation is dose-related (Abou Hammoud, Simon, Urien, et al., 2009; Aubrun, Monsel, Langeron, et al., 2001); therefore, opioid orders and hospital protocols should include the expectation that nurses will stop titration or promptly decrease the opioid dose whenever excessive sedation is detected. For example, the opioid dose should be decreased by 25% to 50% in patients receiving PCA or epidural analgesia who have a sedation level of 3 on the POSS (see Box 19-4 and see Form 17-1 [p. 464] for an example of how decreases in dose can be included in opioid order sets). Monitoring of sedation level and respiratory status is then increased in frequency (e.g., every 15 to 30 minutes) until the sedation level is less than 3 and the respiratory status is stable (see p. 517 for respiratory assessment). Nonsedating nonopioid analgesics can be added or increased to facilitate opioid dose reduction. Routine administration of these agents as part of a multimodal approach initiated preoperatively or as soon as the patient is started on opioid therapy (e.g., admission to the PACU) is essential to help prevent excessive sedation from occurring later in the course of care. Giving less opioid more frequently to decrease peak concentration may be effective in reducing sedation in some cases. For example, large PCA doses and long lockout (delay) intervals can be reduced in patients who report excessive sedation following PCA injection. Catheter migration to either the intravascular or intrathecal space should be ruled out in patients receiving epidural analgesia who have a sudden change in sedation level (see Box 15-4 on p. 433). See the following patient example and Table 19-1 for treatment of opioid-induced sedation during short-term opioid therapy and the following section on prevention and treatment of clinically significant respiratory depression.

Respiratory Depression During Long-Term Opioid Therapy

Clinically significant respiratory depression is the most feared of the opioid-induced adverse effects, but like sedation, tolerance to respiratory depression develops over a period of days to weeks and rarely occurs if opioids are titrated according to patient response (Hanks, Cherny, Fallon, 2004). The longer the patient receives opioids, the wider the margin of safety. Therefore, fear of respiratory depression in patients who have been receiving opioids for more than 1 week should not pose a barrier to administering adequate opioid doses (see later discussion of the postoperative opioid tolerant patient). Similar to opioid-naïve patients, when respiratory depression occurs in an opioid-tolerant individual, it is preceded by increased sedation; however, if appropriate steps are taken to address persistent sedation, respiratory depression is unusual in patients receiving long-term opioid therapy (Hanks, Cherny, Fallon, 2004).

Respiratory Depression in Opioid-Naïve Patients

The incidence of opioid-induced respiratory depression in opioid-naïve patients is unclear, mainly because of variations in the definition of respiratory depression used in research (Cashman, Dolin, 2004; Dahan, Aarts, Smith, 2010). This has resulted in reported incidences based on a variety of indicators, including respiratory rate, number of hypoxemic or hypercarbic episodes, and the need for treatment with naloxone (rescue) (Cashman, Dolin, 2004). The incidence of respiratory depression may be much higher than reported here because of the lack of consensus on definitions and uniform reporting as well as the reliance on retrospective data, the source of most published reports, which may not record the reasons for interventions (Etches, 1994).

Incidence of respiratory depression varies according to the method of opioid administration. A small study (N = 38) reported that no episodes of respiratory depression occurred in patients who were randomized to receive 10 mg of morphine by the IV or the IM route of administration (Tveita, Thoner, Klepstad, et al., 2008). A systematic review of IV PCA reported a range of respiratory depression of 0.6% to 15.2% when hypoventilation and oxygen desaturation were used as indicators (Walder, Schafer, Henzi, et al., 2001). Another systematic review of the incidences of respiratory depression associated with IM, IV PCA, and epidural opioids found wide variations in indicators resulting in wide ranges of incidences (Cashman, Dolin, 2004). Using hypoventilation and oxygen desaturation as indicators, the researchers reported mean incidences of between 0.8% to 37.0% (IM), 1.2% to 11.5% (IV PCA), and 1.1% to 15.1% (epidural). They concluded that an incidence of 1% or less (defined by a low respiratory rate) can be expected in patients receiving IV PCA or epidural analgesia and cared for by an acute pain service. They noted that if oxygen desaturation is used as a definer, a much higher percentage should be expected (see the discussion of respiratory assessment later in the chapter). In a review of the literature, Dahan and colleagues state that the many studies that have been conducted on patients receiving IV PCA yield an incidence of respiratory depression between 0.2% and 0.5% (Dahan, Aarts, Smith, 2010); however, it has been pointed out that many of the studies of IV PCA were published in the early 1990s before emphasis was placed on reducing patients’ pain intensity ratings, and the incidence of oversedation and respiratory depression may actually be much higher today (Taylor, Voytovich, Kozol, 2003).

A prospective study of nearly 3000 patients receiving either PCEA morphine with basal rate (N = 1670) or IV PCA morphine with basal rate (N = 1026) after major surgery found a higher incidence of respiratory depression (defined as less than 8 breaths/min) in the IV PCA group (1.2%) than in the epidural group (0.04%) (Flisberg, Rudin, Linner, et al., 2003) (see Chapter 17 for safe use of a basal rate with IV PCA). In their neuraxial analgesia practice guidelines, the American Society of Anesthesiologists (ASA) reported an incidence of 0.01% to 7% with single-bolus intrathecal morphine and 0.08% to 3% with single-bolus epidural morphine (ASA Task Force on Neuraxial Opioids, 2009). They suggested that the range of incidence for respiratory depression with single-bolus neuraxial opioids (0.01% to 7%) is similar to that of IM, IV, and IV PCA opioids and recommended continuous epidural opioids over parenteral opioids as a way of reducing the risk of respiratory depression.

A significantly higher incidence of respiratory depression than had been previously reported was found when continuous pulse oximetry (O₂ saturation) and capnography (ETCO₂) were utilized to evaluate respiratory depression in a study of 178 postoperative patients receiving morphine or meperidine by IV PCA (Overdyk, Carter, Maddox, et al., 2007). The incidence of oxygen desaturation (less than 90%) lasting 2 or more minutes was 21.4% and lasting 3 or more minutes was 11.8%; the incidence of bradypnea (rate less than 10 breaths/min) lasting 2 or more minutes was 58.4% and lasting 3 or more minutes was 41%. Only 1 patient (0.6%) required naloxone treatment. The researchers attributed the low need for naloxone rescue to the monitoring equipment’s audible alarm that arouses a patient and summons a nurse, but they suggested that more research is needed to verify this (see p. 518 for a discussion of mechanical monitoring). An interesting finding in this study was that despite consuming approximately two times more opioid, patients receiving a basal rate with IV PCA had a significantly lower incidence of bradypnea (32% vs. 53%) and desaturation (8% vs. 17%) than those who did not have a basal rate. The reason for this was not clear.

Reversing Respiratory Depression: Naloxone Administration

The opioid antagonist naloxone has been used since the early 1970s to reverse respiratory depression related to opioid overdose (Buck, 2002). It is administered IV and has a rapid onset of action (2 minutes) with a peak concentration in 10 minutes, duration of 1 to 4 hours, and a half-life of 30 to 81 minutes (Leavitt, 2009). The extent and duration of naloxone reversal of opioid effects varies and is dependent on many factors, such as specific opioid used, opioid dose, method of administration, and concurrent medications (Dahan, Aarts, Smith, 2010).

If it is necessary to use naloxone to reverse clinically significant respiratory depression, it should be titrated very carefully (APS, 2003). Sometimes more than one dose of naloxone is necessary because naloxone has a shorter duration (1 hour in most patients) than most opioids (Dahan, Aarts, Smith, 2010); however, giving too much naloxone or giving it too fast can precipitate severe pain that is extremely difficult to control and increase sympathetic activity leading to hypertension, tachycardia, ventricular dysrhythmias, pulmonary edema, and cardiac arrest (Brimacombe, Archdeacon, Newell, et al., 1991; O’Malley-Dafner, Davies, 2000). Box 19-6 outlines the procedure for correctly administering naloxone to an adult. Hospital protocols and opioid orders should include the expectation that nurses will administer naloxone in accordance with the procedure outlined in Box 19-6 whenever a patient is found to have clinically significant opioid-induced respiratory depression.

In physically dependent patients, withdrawal syndrome can be precipitated by naloxone administration; patients who have been receiving opioids for more than 1 week may be exquisitely sensitive to antagonists (APS, 2003). If sedation and respiratory depression occurs during administration of transdermal fentanyl, the patch should be removed immediately, and if naloxone is necessary, treatment will be needed for a prolonged period and the typical approach involves a naloxone infusion over a 12- to 24-hour period (e.g., mix 0.8 mg of naloxone in 250 mL of IV solution and run at 9 to 20 mL/h to administer approximately 0.2 to 0.6 mg/h naloxone). The patient must be closely monitored for at least 24 hours after discontinuation of the transdermal fentanyl.

IV access for naloxone administration usually is maintained for 24 hours after a stable epidural opioid dose has been achieved. After that time, naloxone rarely is needed because most patients are alert and, in many cases, are ready for a decrease in the epidural opioid dose.

Although no formulation is commercially available in the United States, naloxone has been administered intranasally. One randomized controlled trial found an equally safe and favorable response within 10 minutes of administering 2 mg intranasally or intramuscularly in the prehospital setting for suspected heroin overdose (Kerr, Kelly, Dietze, et al., 2009). A retrospective study compared intranasal naloxone with IV naloxone for opioid overdose in the prehospital setting and found the mean time to clinical response was longer in those who received intranasal naloxone, but the researchers suggested that the response time was equivalent when the time required to prepare an IV injection of naloxone is considered (Robertson, Hendey, Stroh, et al., 2009). In both of these studies, patients who received intranasal naloxone were more likely to require an additional rescue dose of naloxone.