Child-Related Considerations

Acetaminophen

Acetaminophen is the most common antipyretic and nonopioid analgesic used in children. It exerts its analgesic effects by blocking central and peripheral prostaglandin synthesis, reducing substance P–induced hyperalgesia, and modulating the production of hyperalgesic nitric oxide in the spinal cord.^119–121 In addition, it has been suggested that acetaminophen produces analgesia via activation of descending serotonergic pathways.^122–124 However, it is likely that its primary site of action may be inhibition of prostaglandin H₂ synthetase at the peroxidase site.¹²³ Effective analgesic and antipyretic effects have been described with plasma concentrations of 5 to 20 µg/mL^125–129; a target effect-site concentration of 10 µg/mL reduces pain after tonsillectomy by 3.6/10 pain units.¹³⁰ The total daily dose of acetaminophen via any route is age- and weight-based but should not exceed 75 mg/kg for children; term and preterm infants require further downward dosing adjustment (60 mg/kg and 45 mg/kg, respectively).

The recommended dose for oral administration is 10 to 15 mg/kg every 4 hours. Acetaminophen has a wide margin of safety when administered in the recommended therapeutic dose range. However, hepatotoxicity has been reported with doses only slightly above the recommended 10- to 15-mg/kg/dose orally for a total of five doses or 75 mg/kg/day, suggesting that acetaminophen may have a narrow therapeutic index in some children.^131,132 Because of these reports and on the advice of a U.S. Food and Drug Administration panel, the manufacturers have reduced the maximum single dose of oral acetaminophen in adults to 650 mg and the maximum daily dose to 3 grams. Acetaminophen is available in a wide variety of formulations, alone or in combination with decongestants, for oral use in a variety of cold remedies, and with opioids for the treatment of moderate to severe pain. There are currently more than 600 over-the-counter acetaminophen-containing products, increasing the risk of an overdose because children may take more than one formulation that contains the drug. Frequent review of medications and parental education is needed to minimize the risk of overdose. In the past, pediatric liquid formulations of acetaminophen as in infant drops were commonly supplied in larger concentrations than that in elixirs, resulting in dosing errors. The current recommendation is to standardize liquid formulations to a single concentration of 32 mg/mL. Acetaminophen can be given orally before surgery; both gastric fluid volume and pH are unchanged after acetaminophen was administered orally 90 minutes before induction of anesthesia.¹³³

Slow and unpredictable absorption of acetaminophen after rectal administration results in variable blood concentrations, with peak concentrations reached between 60 and 180 minutes after administration.^128,134,135 There is a dose-response relationship for rectal acetaminophen. The morphine-sparing effects of 40 mg/kg and 60 mg/kg of rectal acetaminophen were greater than those of 20 mg/kg and placebo in children undergoing ambulatory surgery.¹³⁶ In children undergoing orthopedic surgery, a loading dose of 40 mg/kg rectal acetaminophen followed by 20 mg/kg every 6 hours yielded serum concentrations of 10 to 20 µg/mL, with no evidence of accumulation over a 24-hour period.¹³⁴ This dosing scheme is now the one most commonly recommended when the rectal route is employed.

Intravenous (IV) formulations of paracetamol (acetaminophen) and its prodrug propacetamol have been used in Europe and Australia for several years and are now available in the United States. IV acetaminophen is available as a 10 mg/mL solution and should be infused over at least 15 minutes in a dose of 10 mg/kg, with a total daily dose not to exceed 75 mg/kg. The maximum dose, regardless of weight, is 750 mg every 6 hours (3 grams per day).

After IV administration of acetaminophen, analgesic onset occurs in 15 minutes and that of antipyresis in 30 minutes.^137,138 IV paracetamol rapidly penetrates the blood-brain barrier in children, yielding detectable concentrations in the cerebrospinal fluid (CSF) within 5 minutes of administration, and peak CSF concentrations within 57 minutes after injection (compared with 2 to 3 hours after rectal or oral administration), thus explaining the fast onset of its analgesic and antipyretic effects.¹³⁹ A large multicenter trial reported that 1 gram of IV paracetamol and 2 grams of IV propacetamol (equivalent to 1 gram acetaminophen) provided superior analgesia with a reduced need for morphine compared with placebo in adults after lower extremity joint replacement.¹⁴⁰ The propacetamol group experienced a greater incidence of local skin reactions and pain on injection compared with the IV paracetamol group. A controlled randomized trial reported that both rectal acetaminophen, 40 mg/kg, and IV acetaminophen, 15 mg/kg, administered after induction of anesthesia in children undergoing adenotonsillectomy, provided good analgesia for the first 6 hours after surgery.¹⁴¹ However, children who received acetaminophen rectally had a greater duration of analgesia and did not require rescue analgesia as early as those in the IV group.¹⁴¹ This is attributable to the slow absorption of rectal acetaminophen causing sustained effective concentrations. A prospective randomized trial comparing rectal acetaminophen to IV propacetamol in infants after craniofacial surgery reported that the IV formulation provided superior analgesia,¹⁴² in part because of reduced bioavailability of acetaminophen by the rectal route.

Another controlled randomized study compared the analgesic efficacy and side effects of fentanyl-placebo versus fentanyl-acetaminophen administered via PCA in 6- to 24-month-old children undergoing ureteroneocystostomy.¹⁴³ Children in the acetaminophen group required significantly less fentanyl and demonstrated a reduced incidence of vomiting and excessive sedation compared with those in the placebo group. Lastly, a large retrospective study reported no differences in alanine transaminase and γ-glutamyl transferase concentrations, and a progressive decrease in aspartate aminotransferase levels, in term and preterm neonates before, during and after IV acetaminophen injection.¹⁴⁴

Nonsteroidal Antiinflammatory Drugs

NSAIDs provide excellent analgesia for mild to moderate pain resulting from surgery, injury, and disease. Their principle mechanism of action is via inhibition of the enzyme prostaglandin H₂ synthetase at the COX site, causing a reduction in the production of prostaglandins at the site of tissue injury, and attenuation of the inflammatory cascade. In addition to their peripheral effects, the NSAIDs have also been shown to exert a direct spinal action by blocking the hyperalgesic response induced by activation of spinal glutamate and substance P receptors.¹⁴⁵ Decreased production of leukotrienes, activation of serotonin pathways, and inhibition of excitatory amino acids, NMDA-mediated hyperalgesia, and central inhibition of prostaglandin biosynthesis have been proposed as additional mechanisms of action.^146,147 The COX-1 enzyme is present in the brain, gastrointestinal tract, kidneys, and platelets and is expressed constitutively. It preserves gastric mucosal integrity and function, platelet aggregation, and renal perfusion. COX-2 expression is induced by inflammation or tissue injury. Selective COX-2 inhibitors reduce inflammation but have less effect on gastric mucosal function and have fewer effects on platelet aggregation, thereby resulting in fewer adverse effects. Their deleterious effects on renal perfusion, however, are no different than the nonselective COX drugs, because COX-2 is constitutively expressed in renal tissues and may be involved in prostaglandin-dependent renal homeostatic processes.¹⁴⁸ The risks of renal toxicity increase in the presence of hypovolemia, cardiac failure, preexisting renal dysfunction, or with the concurrent use of other nephrotoxic drugs. Reports of thrombotic cardiovascular and CNS events after both long-term and short-term use in adults led to withdrawal of two of the COX-2 inhibitors, rofecoxib and valdecoxib from the market.^149,150 Similar data are unavailable to date in children; consequently, the risk of these agents causing thrombotic complications in children remains unknown. Most pediatric studies have evaluated the use of nonselective COX medications. In adult studies, COX-2 inhibitors have generally, but not always, produced analgesia roughly equivalent to that of traditional NSAIDs. Ibuprofen, one of the oldest orally administered NSAIDs, has been used extensively for treatment of fever and pain related to surgery, trauma, arthritis, menstrual cramps, and sickle cell disease. A large, controlled, randomized, double-blind study reported a greater decrease in VAS pain scores with ibuprofen than with acetaminophen or codeine in children presenting to the emergency department with acute pain after musculoskeletal trauma.¹⁵¹ Additionally, more children who received ibuprofen had VAS scores less than 30 on a 0-to-100-mm VAS scale than in the other two groups. The recommended dose of ibuprofen is 6 to 10 mg/kg every 6 hours. Like acetaminophen, ibuprofen is available in a variety of formulations and concentrations, placing children at risk for an overdose. For pediatric use, ibuprofen is available as:

Diclofenac provides effective analgesia after minor surgical procedures in children. It is available only as an oral tablet in the United States, but it is available as a suppository and in the injectable form in several countries. The pediatric dose of diclofenac is 1 mg/kg every 8 hours orally, 0.5 mg/kg rectally, and 0.3 mg/kg IV.¹⁵² The oral and rectal doses reflect bioavailabilities of 0.36, 0.35, and 0.6 for suspension, dispersible tablets, and suppository, respectively. When diclofenac was administered rectally, the relative bioavailability was greater and the peak concentration was reached earlier than after enteric coated tablets administered orally.¹⁵³ Children who received diclofenac experienced comparable analgesia to those who received caudal bupivacaine or IV ketorolac for inguinal hernia repair.^154–156 In children undergoing tonsillectomy and/or adenoidectomy, diclofenac yielded superior analgesia with less supplemental opioid dosing, less nausea and vomiting, and earlier resumption of oral intake compared with acetaminophen.^157,158 Although there are occasional reports of increased bleeding and restlessness in the recovery room in children who received diclofenac compared with those who had received papaveretum during tonsillectomy,¹⁵⁹ a Cochrane review established that NSAIDs did not cause any increase in bleeding that required a return to the operating room (OR) for children. There was significantly less nausea and vomiting with NSAIDs compared with alternative analgesics, suggesting their benefits outweigh their negative aspects.¹⁶⁰

Ketorolac, indomethacin and ibuprofen are the only injectable NSAIDs available in the United States. Indomethacin is the only NSAID used for closure of patent ductus arteriosus in preterm neonates. The IV formulation of ibuprofen is only labeled for adults in the United States. Clinical trials are currently under way in children. Ketoprofen and diclofenac are other injectable NSAIDs that are available outside the United States. A large multicenter study compared the risks of serious adverse events from IV ketorolac, ketoprofen, and diclofenac in more than 11,000 adults undergoing major surgery.¹⁶¹ The results indicated that 1.4% of adults experienced a serious adverse outcome, including surgical site bleeding (1%), death (0.17%), severe allergic reactions (0.12%), renal failure (0.09%), and gastrointestinal bleeding (0.04%), with no differences in outcomes among the groups; similar large-scale studies are not available for children.

Ketorolac has been shown to provide postoperative analgesia similar to opioids, in children of all ages.^162–165 Its benefits include lack of opioid adverse effects (respiratory depression, sedation, nausea, and pruritus) making it an attractive choice for the treatment of postoperative pain. However, in common with all NSAIDs, it carries risks of platelet dysfunction, gastrointestinal bleeding, and renal dysfunction. Ketorolac (1 mg/kg) given to 18 preterm and term neonates undergoing painful procedures in the OR or the neonatal intensive care unit,¹⁶³ revealed reduced pain scores (Neonatal Infant Pain Scale) with no incidents of systemic or local bleeding and no hematologic, hepatic, or renal complications (note that this dose is twice the usually recommended dose of 0.5 mg/kg). Similarly, no adverse effects on surgical drain output, renal or hepatic function tests, or oxygen saturation after major surgery were noted in 37 infants and toddlers between 6 and 18 months of age.¹⁶⁶ Children in that study received continuous morphine infusions postoperatively, confounding the evaluation of the analgesic efficacy of ketorolac. Finally, ketorolac has been used to supplement opioid analgesia, with no increase in renal or bleeding complications in infants and children after open heart surgery.^167–169 Nevertheless, because ketorolac can reduce renal blood flow, many recommend that its course be limited to 48 to 72 hours, and that renal function be checked if a course of administration greater than 72 hours is required. In single dose studies, the pharmacokinetics (PK) of ketorolac in infants less than 12 months of age appear to be homogeneous, although there was a trend toward reduced clearance in the infants less than 6 months of age.¹⁷⁰

In an effort to avoid the respiratory depressant effects of opioids after airway surgery, several studies investigated the safety and benefits of ketorolac in children undergoing tonsillectomy.^171–175 These early reports of ketorolac adversely skewed analyses of the adverse effects of NSAIDs after tonsillectomy. All but one of these studies¹⁷¹ found a two- to fivefold increase in bleeding complications, including measured blood loss, ease of achieving hemostasis, and bleeding episodes in the postanesthesia care unit (PACU), necessitating reexploration and hospital admission in some cases. Two of these studies were terminated prematurely, when preliminary data showed an unacceptably greater risk of bleeding in children who had received ketorolac.^172,175 In one of these two studies, ketorolac was given at the end of surgery after hemostasis had been achieved.¹⁷² The benefits of ketorolac, including adequacy of analgesia, resumption of oral intake, and reduction in nausea, vomiting, and sedation were modest. This issue is further confounded by conflicting results yielded by two meta-analyses that evaluated the related literature. In one of these, the use of aspirin, but not of NSAIDs (diclofenac or ibuprofen), significantly increased the risk of post-tonsillectomy hemorrhage compared with either acetaminophen with codeine or tramadol for postoperative analgesia.¹⁷⁶ The other reported no increase in intraoperative blood loss, postoperative bleeding, or admission because of bleeding, but did show a statistically significant increase in the rate of reoperation for bleeding in children who received an NSAID in the postoperative period compared to those who did not.¹⁷⁷ Taken together, these data suggest that the use of NSAIDs during or after tonsillectomy is best avoided, and alternative analgesics, such as acetaminophen and tramadol, be considered to reduce opioid requirements. A large multicenter study in adults found that the risk of gastrointestinal and operative site bleeding associated with ketorolac was larger and clinically important when ketorolac was used in larger doses, in older subjects, and for more than 5 days.¹⁷⁸

Another contentious issue regarding NSAIDs relates to their effects on bone healing and their use in children undergoing spinal fusion. Prostaglandins play an integral role in bone metabolism and significantly influence bone resorption and formation; however, their effects on bone formation predominate. NSAIDs inhibit the formation of prostaglandins, thereby raising the concern that they could promote nonunion after spinal fusion. Studies in rabbits and some studies in adults have reported a greater incidence of nonunion or pseudarthrosis, particularly with the use of large doses of ketorolac.^179,180 However, no differences in curve progression, hardware failure, pseudarthrosis, or need for reoperation have been found in children and adolescents who received ketorolac in the immediate postoperative period compared with those who did not.^181–183 Of note, the majority of the pediatric data are from otherwise healthy children with idiopathic scoliosis, making it problematic to extrapolate these data to children with comorbidities or those with neuromuscular scoliosis. There is no unique advantage of the IV route with NSAIDs. There is also no evidence that IV ketorolac is a more potent analgesic than comparable (i.e., equipotent) doses of a number of other NSAIDs, administered by oral or rectal routes.¹⁸⁴

A recent meta-analysis of the use of NSAIDs for postoperative pain included 27 studies and compared 567 children who received NSAIDs to 418 children who did not.¹⁸⁵ This study found that coadministration of NSAIDs and opioids during the perioperative period decreased opioid requirement in the PACU and the first 24 hours after surgery, decreased pain intensity in the PACU, and postoperative nausea and vomiting (PONV) during the first 24 hours postoperatively. Additionally, coadministration of acetaminophen with NSAIDS and opioids reduced pain intensity for the first 24 hours postoperatively. Other investigators demonstrated up to a 30% opioid-sparing effect in children who receive acetaminophen and diclofenac in addition to PCA. Therefore, in the absence of contraindications, it has been recommended that NSAIDs be used as part of a multimodal regimen to manage postoperative pain and to decrease opioid consumption in children.¹⁸⁵

Tramadol

Tramadol is a synthetic analogue of codeine that exerts its analgesic properties by two complementary mechanisms. One of its metabolites has a weak affinity for the µ opioid receptor with no affinity for the δ or the κ receptors. In addition to its mild opioid effects, it also inhibits serotonin and norepinephrine uptake. Its main advantages over opioids include reduced incidences of respiratory depression, sedation, nausea, and vomiting. Additionally, because it does not inhibit prostaglandin synthesis, it does not cause the adverse effects commonly reported with NSAIDs, including peptic ulceration and renal and platelet dysfunction. Adverse effects associated with its use include nausea and vomiting (9% to 10% of cases), pruritus (7%), and rash (4%).¹⁸⁶ It is known to cause dizziness and its use has been associated with seizures. Tramadol is available only in tablet form alone or in combination with acetaminophen in the United States. However, it is available in a liquid formulation (and as oral drops for infants), as a suppository, and as an injectable solution in other countries, allowing for greater flexibility of dosing. Therefore, it has been used to provide analgesia by a number of routes, including oral, rectal, IV (including PCA devices), into the caudal epidural space, and by local infiltration.

Tramadol is used for postoperative pain treatment in children undergoing ambulatory surgery and has also been used when transitioning from IV opioids to oral analgesics. Two doses of tramadol (1 mg/kg and 2 mg/kg orally) were compared in children who were being transitioned from morphine PCA. Children who received 2 mg/kg required fewer supplemental analgesics with no difference in adverse effects compared with those who had received 1 mg/kg.¹⁸⁶ Tramadol, 2 mg/kg IV, produced similar analgesia and sedation, with fewer episodes of oxygen desaturation compared with morphine 0.1 mg/kg IV, in children with obstructive sleep apnea undergoing adenotonsillectomy.¹⁸⁷ Tramadol has also been found to produce a similar analgesic effect as that of ilioinguinal and iliohypogastric nerve blocks in children undergoing herniorraphy.¹⁸⁸ The tramadol group, however, experienced a greater incidence of nausea and vomiting. Tramadol PCA has also been found to provide adequate analgesia with less sedation, earlier awakening, and earlier extubation in children undergoing atrial or ventricular septal defect repair compared with those who received morphine via PCA.¹⁸⁹

Tramadol has also been effective when administered via the neuraxial space. Caudal tramadol (2 mg/kg) produced reliable postoperative analgesia comparable to that produced by caudal morphine (30 µg/kg) in children undergoing inguinal hernia repair.¹⁹⁰ No additional pain medications were required in the first 24 hours in more than 90% of children in each group. Rigorous drug-specific neurotoxicity studies, however, are lacking. Another study compared the analgesic efficacy of 2 mg/kg tramadol administered IV or by peritonsillar infiltration in children undergoing adenotonsillectomy.¹⁹¹ Both groups experienced excellent analgesia in the first hour. However, the local infiltration group experienced more prolonged analgesia and required fewer rescue doses of acetaminophen compared with the IV group. Overall, tramadol appears to be an analgesic of medium potency with a low incidence of adverse effects that may be used alone for mild to moderate pain and for its opioid-sparing effect in children with severe pain.

Ketamine

There has been increasing interest in the use of ketamine, an NMDA-receptor antagonist, in the treatment of both chronic and acute pain. Its professed benefits include an opioid-sparing effect, avoidance of opioid tolerance, prevention of central sensitization and wind-up, mitigation of opioid-induced hyperalgesia, and provision of synergistic analgesia in multimodal regimens by virtue of its own antinociceptive properties. Case series in children with intractable pain resulting from advanced stages of cancer have reported reduction in opioid requirement, decreased opioid adverse effects, improvement in pain control and function, and increased ability to interact with their families.^192–194

Studies evaluating the use of ketamine alone or in combination with opioids for acute postoperative pain in children have yielded equivocal results. In one study, children undergoing tonsillectomy who received IV ketamine 0.5 mg/kg after induction or at the end of surgery experienced reduced pain scores and required fewer rescue analgesics compared with those who received placebo.¹⁹⁵ All children in this study received a standardized analgesic regimen, including rectal diclofenac before the start of surgery and oral acetaminophen at scheduled intervals postoperatively. Another study reported reduced pain scores and reduced requirement for rescue analgesics in children who received ketamine as a bolus dose and by infusion that began before the start of a tonsillectomy, compared with those who received a single bolus dose of ketamine at the end of surgery.¹⁹⁶ Intramuscular (IM) ketamine 0.5 mg/kg has also produced equivalent analgesia in terms of similar pain scores and need for rescue analgesics compared with IM morphine 0.1 mg/kg as sole analgesics for tonsillectomy.¹⁹⁷ Other studies have found no such benefits when ketamine was compared with placebo for tonsillectomy, urologic, and orthopedic surgery.^198–201

A meta-analysis of 35 randomized controlled trials compared 567 children who received ketamine as an adjuvant analgesic by a variety of routes for a variety of surgical procedures with 418 who did not receive ketamine.²⁰² This study found that, although the use of ketamine was associated with reduced pain intensity in the PACU and a reduced need for nonopioids, it did not demonstrate an opioid-sparing effect. A systematic review of 37 studies included 4 studies in children, two of which demonstrated beneficial effects of ketamine administered as an adjuvant analgesic and two found no benefits.²⁰³ The investigators could draw no conclusions regarding the use of ketamine as an adjuvant analgesic. The use of ketamine in all the above studies was associated with only a few mild and self-limiting adverse effects. Further investigation is needed to evaluate the benefits of low-dose ketamine for acute postoperative pain before its routine use can be recommended.

Delivery Techniques

The blood concentration of opioids must be maintained within a therapeutic range to provide effective analgesia and avoid undesirable adverse effects, such as excessive sedation and respiratory depression. Both the dose and the route by which the opioid is delivered determine how well one is able to maintain the blood concentration within this therapeutic window and minimize adverse effects.

Selection of Opioids for Parenteral Use

Morphine is the opioid most commonly used for postoperative analgesia and has been extensively studied in all pediatric age groups. After major abdominal, thoracic, and orthopedic surgery, children who received continuous morphine infusions had reduced pain scores compared with those who received intermittent IM or IV injections.^222–224 However, other investigators were only able to demonstrate reduced pain scores with continuous morphine infusions compared with intermittent IV injections of morphine in children between 1 and 3 years of age, and not in infants in the first year of life.^225,226 Similarly, evidence of the beneficial effects of opioid analgesia in ameliorating the postoperative response to surgical stress are conflicting. A significant reduction in serum β-endorphin concentrations has been reported in neonates after the initiation of a continuous infusion of morphine in the postoperative period.²²⁷ In neonates whose lungs were mechanically ventilated, both epinephrine and norepinephrine concentrations decreased significantly after the initiation of morphine or fentanyl infusions.²²⁸ However, β-endorphin concentrations decreased only in the children who received fentanyl. When the effects of continuous infusions of morphine were compared with those of intermittent IV injections of morphine on the stress response in children between 1 and 3 years of age, reduced glucose concentrations in the continuous infusion group suggested only a modest ablation of the stress response in this age group.²²⁵

Several studies have described the PK of morphine administered as a continuous infusion and evaluated the pharmacodynamic effects of morphine on respiratory indices in neonates, infants, and children after various surgical procedures. In children 14 months to 17 years of age who underwent cardiac surgery,²²⁹ morphine infusions were adjusted between 10 and 50 µg/kg/hr to minimize discomfort and avoid excessive sedation. Supplemental boluses of 100 µg/kg morphine were administered for breakthrough pain. Steady-state morphine concentrations were achieved in 4 hours. Those children who could self-report their pain reported good analgesia with morphine concentrations in excess of 12 ng/mL. Morphine infusions of 10 to 30 µg/kg/hr yielded mean serum concentrations between 10 and 22 ng/mL with less than 2% experiencing evidence of respiratory depression (Paco₂ greater than 50 mm Hg). Furthermore, children who received morphine infusions of 10 to 30 µg/kg/hr breathed spontaneously after extubation of the trachea, and those who were weaned from assisted to spontaneous ventilation maintained a normal Paco₂. On the other hand, 60% (3 of 5 children) who received a greater infusion rate of morphine, 40 to 50 µg/kg/hr, experienced hypercarbia (Paco₂ 48 to 66 mm Hg). A subsequent study by the same investigators evaluated the severity of respiratory depression in infants and children aged 2 days to 18 months treated with morphine. Of those whose morphine concentrations exceeded 20 ng/mL, approximately 70% experienced respiratory depression (Paco₂ greater than 55 mm Hg and/or a depressed slope of the CO₂ response curve) compared with 15% to 28% of those whose concentrations were less than 20 ng/mL.²³⁰ The investigators suggested a steady-state morphine concentration of 20 ng/mL as a threshold concentration for respiratory depression in this age group.

Previous studies determined that the clearance of morphine is impaired in preterm infants and that clearance increases with postconception age.²³¹ Additionally, morphine clearance is impaired in full-term infants up to 1 to 2 months of age, at which time it is comparable with that in older children and adults.^98,232 Preterm and full-term neonates, therefore, have a narrower therapeutic window for morphine analgesia compared with older children. Indeed, these groups have reduced morphine requirements postoperatively, requiring fewer rescue doses of morphine when receiving continuous infusions or intermittent bolus doses.²³³ Therefore, opioids should be carefully titrated in infants in a monitored environment with significantly reduced continuous infusion rates. Based on PK modeling and morphine clearance predictions, a target morphine concentration of 10 ng/mL can be achieved with morphine infusions ranging from 5 µg/kg/hr in term neonates to 16 µg/kg/hr in 1- to 3-year-old children (see Chapter 6).²³⁴

Pharmacodynamic differences between infants and children have been postulated as the mechanism responsible for the greater sensitivity of infants (compared with older children) to the respiratory depressant effects of opioids. However, this may not be the case. Although rodent data suggest that the brain concentrations of opioids in neonates are greater than those in older children at similar serum concentrations,²³⁵ these findings may not be applicable to humans. Neonatal rats have a relatively immature brain and a far more permeable blood-brain barrier than that in human infants. Consequently, the rodent may not be an appropriate model to depict the human condition.²³⁶ It appears that the “increased sensitivity” is related, at least in part, to PK variables, perhaps in some measure as a result of a neonate’s decreased conjugating ability.

Regardless of the mechanism, respiratory depression remains the most feared adverse effect of opioids administered by any route. Neonates and infants younger than 6 months of age are at greater risk for opioid-induced respiratory depression because the ventilatory responses to airway obstruction, hypoxemia, and hypercapnia are immature at birth and mature over the first several months of life in preterm as well as full-term infants (see also Figs. 4-8 and 4-9). Indeed, there was a 4.5% incidence of failure to wean from the ventilator and a 13.5% incidence of apnea (30 seconds or more that required intervention) or severe respiratory depression in spontaneously breathing neonates who received opioids for postoperative pain.²³⁷ Another report of a 3-year surveillance period for adverse drug reactions described 15 children aged 2 days to 17 years who experienced opioid-induced respiratory depression.²³⁸ Respiratory depression in the latter study was defined as apnea, hypoxemia, cyanosis, a marked decrease in respiratory rate, or a need for naloxone. Although this study was unable to define the incidence of respiratory depression because the denominator was unknown, it did identify several predisposing factors, including age younger than 1 year (7 of 15 children), drug errors (including prescription and administration errors; 6 of 15 children), concurrent medical problems (diminished respiratory reserve, hepatic, and/or renal impairment), and concurrent sedative drugs. The prospective UK audit found 14 cases of respiratory depression (out of 10,726 total infusions, or 0.13%), 10 with nurse-controlled anesthesia (NCA), 2 with continuous infusions, and 2 with PCA.²²⁰ Potentially contributing risk factors in half of the cases included very young age and neurodevelopmental, respiratory, or cardiac disease. In contrast to the above studies, no case of respiratory depression was reported in 110 children older than 3 months of age who received opioid infusions postoperatively.²³⁹ Interpretation of this literature is confounded by different monitoring techniques and different definitions of respiratory depression. For instance, in the latter study, a 4.5% incidence of clinically significant hypoxemia was reported but was not included in their definition of respiratory depression. Additionally, children in that study were monitored with hourly documentation of respiratory rate, but oxygen saturation was not monitored after discharge from PACU, thereby reducing their ability to detect the more subtle episodes of respiratory depression. In summary, the results of these studies suggest that children who receive opioids require careful monitoring for respiratory depression, with appropriate age-based reduction of dosage, particularly for neonates and infants younger than 6 months of age.

The most common adverse effect of opioid therapy is nausea and vomiting. One study reported nausea and vomiting in 34 of 80 children (42.5%) who received postoperative morphine infusions. These were well managed with antiemetic therapy in all but 2 children who required discontinuation of the opioids.²³⁹ In the same study, the incidence of pruritus and urinary retention were both 13% and that of dysphoria was 7%. Seizures have been reported in two neonates who had received bolus doses of morphine followed by infusions of 32 and 40 µg/kg/hr and whose serum morphine concentrations were 61 and 90 ng/mL, respectively.²⁴⁰ Irregular jerking movements, as well as one case of a generalized seizure, have been reported in children 1 to 15 years of age receiving postoperative morphine infusions.²²² Metoclopramide, 0.10 to 0.15 mg/kg (100 to 150 µg/kg) given IV, is an effective antiemetic but may also cause sedation and dystonia. The serotonin-receptor antagonist antiemetics, such as ondansetron and dolasetron, have the advantage of virtually eliminating the risk of dystonic or oculogyric reactions that occur with phenothiazines, butyrophenones, and metoclopramide. However, headaches occur in a small number of those who receive serotonin-receptor antagonists. A “microdose” naloxone infusion (0.25 to 1.0 µg/kg/hr) reverses the incidence of both nausea and pruritus after opioids without affecting the analgesia or opioid consumption.^241,242 A more recent dose-escalation study demonstrated that doses of 1 to 1.65 µg/kg/hr resulted in greater efficacy in reducing side effects, particularly pruritus, without degrading analgesia.²⁴² It is likely that these results may be generalized to other routes of opioid administration.

Opioid-induced bowel dysfunction reported in more than 90% of patients on opioid therapy, occurs by blocking propulsive peristalsis, inhibiting secretion and increasing reabsorption of intestinal fluids, and decreasing the activity of excitatory and inhibitory neurons in the myenteric plexus. Bowel dysfunction manifests as abdominal distension and bloating, delayed gastric emptying, and constipation. Aggressive prophylactic measures, including osmotic, lubricant, or stimulant laxatives, should be prescribed early in the course of treatment. A newer selective gastrointestinal peripheral µ-opioid receptor antagonist, methylnaltrexone, was approved for use in adults in 2008. Although adult studies have shown promising results with the use of this agent,^243,244 its pediatric use has been described in only one case report.²⁴⁵ A neonate who experienced a severe ileus during fentanyl administration after major abdominal surgery was noted to have resolution of the signs within 15 minutes of IV methylnaltrexone (0.15 mg/kg). She received 5 daily doses of methylnaltrexone without reversal of analgesia or occurrence of withdrawal.²⁴⁵

Fentanyl may be a useful substitute for morphine in children with hemodynamic instability, in whom a decrease in peripheral vascular tone is undesirable, and in whom histamine release caused by morphine is not well-tolerated. Additionally, its rapid onset of analgesia makes it ideal for children with severe escalating pain who require urgent pain relief. Fentanyl is metabolized by the liver into an inactive metabolite, norfentanyl, which is excreted via the kidneys. It is 80 to 100 times more potent than morphine. Although its elimination half-life is significantly less than that for morphine, its context-sensitive half-life during chronic infusion increases exponentially as a result of growing tissue storage (see Chapter 6). Like morphine, the elimination half-life of fentanyl in neonates is nearly twice that in adults, predisposing them to a greater risk for accumulation compared with older infants.^246,247 As with morphine, a reduction in hepatic blood flow in very young infants further decreases fentanyl conjugation. For a given bolus of fentanyl, plasma concentrations in infants between 3 months and 1 year of age are less than those in older children and adults.²⁴⁸ This finding is consistent with the almost twofold greater clearance of fentanyl in children compared with neonates. In children 18 days to 14 years of age who were mechanically ventilated, the clearance of fentanyl was age related yet quite variable, with the slowest clearance occurring in infants younger than 6 months of age and the most rapid in those between 6 months and 6 years of age.²⁴⁹ The clearance of fentanyl is slow in preterm infants, with the clearance correlating with the postnatal age.²⁵⁰

Fentanyl is known to cause all of the adverse effects reported with opioids, including pruritus, nausea, vomiting, constipation, and sedation. Respiratory depression and chest wall and glottic rigidity, however, are its most feared adverse effects. One study compared the incidence of respiratory depression in full-term and former preterm infants and young infants receiving 2 µg/kg bolus doses of fentanyl every 2 hours or a continuous infusion of 1 µg/kg/hr after abdominal or thoracic surgery.²⁵¹ Randomization was terminated prematurely because of a sixfold greater incidence of apnea that required intervention in the bolus dose group compared with the continuous infusion group (89% vs. 14%). The continuous infusion arm was continued for another 20 children, resulting in a 25% incidence of apnea in this group. In contrast, the incidence of respiratory depression (based on transcutaneous Paco₂ measurements and the incidence of apnea) for a given plasma fentanyl concentration in infants 1 to 12 months of age and children 1 to 5 years of age was less than that in adults undergoing hernia repair or other peripheral surgery.²⁵² Differences in surgical procedures and the inclusion of preterm infants in the former study may account for the significant difference in the incidence of apnea found in these two studies.

Although chest wall rigidity usually occurs after the rapid bolus administration of high-dose fentanyl, it has also been reported in an infant after a low-dose continuous infusion of fentanyl. Chest wall rigidity was reported in 9% of preterm and full-term neonates who received an average of 4.9 µg/kg over a 2 to 3 minute period for a procedure or for perioperative analgesia.²⁵³ In every case, naloxone reversed the chest wall rigidity. Additionally, a case of chest wall rigidity has been reported in a preterm neonate after high-dose fentanyl was administered to the parturient before a cesarean section.²⁵⁴ Although administration of naloxone has been used successfully to treat cases of chest wall rigidity, severe cases associated with rapid oxygen desaturation may require the use of neuromuscular blocking drugs and mechanical ventilation.

The use of continuous fentanyl infusions in infants and children has been associated with a rapid development of tolerance, as indicated by a steady increase in infusion rate to maintain the desired effect^255,256 and a large incidence of opioid withdrawal syndrome after termination of the infusion.^256,257 The incidence of opioid withdrawal is directly related to the total dose administered and the duration of infusion.^256,257 Neonatal abstinence syndrome has been reported in 21 of 37 neonates (57%) after continuous fentanyl infusions during extracorporeal membrane oxygenation.²⁵⁶ Both a total fentanyl dose in excess of 1.6 mg/kg and extracorporeal membrane oxygenation that lasted more than 5 days were predictors of opioid withdrawal. A similar incidence has been reported in 23 children 1 week to 22 months of age who received continuous fentanyl infusions during mechanical ventilation.²⁵⁷ This study also found that a total dose of 1.5 mg/kg of fentanyl over 5 days was associated with a greater than 50% incidence of withdrawal symptoms. Furthermore, a total dose of 2.5 mg/kg as a continuous infusion over 9 days was 100% predictive of the occurrence of withdrawal. Finally, movement disorder and irritability have been reported after withdrawal of fentanyl infusion in five infants who were mechanically ventilated.²⁵⁸ None of the infants who developed the movement disorder had received another opioid after withdrawal of fentanyl, whereas five of eight controls who did not develop withdrawal during the same period had received a substitute opioid. These data suggest that opioid withdrawal occurs earlier and with greater frequency after fentanyl infusions compared with other opioids. Therefore it seems prudent to use fentanyl infusions for pain relief during periods of hemodynamic instability, such as in the early postoperative period, and to transition to another opioid, such as morphine, as soon as the child is stabilized. Children who require fentanyl infusions for 5 days or more should undergo a slow taper (e.g., 10% decrease every 12 hours) or be transitioned to another parenteral or oral opioid regimen.

Hydromorphone has a spectrum of action similar to that of morphine. Adult opioid equipotency data suggest that it is 3.5 to 7 times as potent as morphine.^259–261 The only pediatric study that was performed in children with mucositis pain after bone marrow transplant reported that a 7 : 1 conversion ratio of morphine to hydromorphone, underestimated hydromorphone requirements by 27%.²⁶² These data suggest that a 5 : 1 conversion ratio may be more appropriate, particularly in children with chronic pain. Despite its widespread use, there are very few studies that evaluated the use of hydromorphone in children. A Cochrane review of studies related to the use of hydromorphone for acute and chronic pain in adults and children found little difference between the analgesic efficacy and adverse effect profile of morphine and hydromorphone.²⁶³ However, several of the studies in this review included small numbers of patients, some were of low quality and only four of them included children. It remains common practice to prescribe a trial of hydromorphone in children who experience unacceptable side effects with morphine.

Meperidine is an opioid that has been used clinically for many years.^264,265 Its potency is approximately one-tenth that of morphine. Accumulation of its active metabolite, normeperidine (which has CNS stimulant properties), after repeated doses of meperidine, places children at risk for seizures.²⁶⁶ Therefore, its use has been restricted to the treatment of postoperative shivering^267,268 or rigors after amphotericin. A single dose of dexmedetomidine (0.5 µg/kg) has been used successfully for the treatment of postoperative shivering and may replace meperidine for this indication.²⁶⁹ Although its short-term use continues by some clinicians for procedural sedation and analgesia, it is preferable to use other analgesics for this purpose. Meperidine is not recommended for PCA or as a continuous infusion in children.

Patient-Controlled Analgesia

PCA was first studied in adults in 1965. The initial interest with this technique was as a research tool for the study of pain. By the early 1970s, it was identified as an excellent strategy for treating pain in the clinical setting, with studies demonstrating that pain relief was achieved by PCA with relatively smaller doses of opioids and with greater patient satisfaction than with conventional methods.²⁷⁰ However, it was not until the late 1980s that PCA was studied in children.²⁷¹ Since that time, it has become the preferred method for opioid delivery in children older than 5 to 6 years of age for acute pain, as well as chronic pain associated with cancer or sickle cell disease.^{262,271–274} The primary benefit of PCA is that it allows children to titrate the analgesic to the extent of their pain. The goal is for the child to self-regulate a blood opioid concentration within the therapeutic range. Most children strike a balance between adequate pain relief on the one hand and adverse effects of the drug on the other. This approach, which grants the child some degree of autonomy, is the rationale given for the fact that pain is an entirely subjective and individual experience and that opioid metabolism and pain perception are quite variable among individuals. It also reduces the apprehension that older children and adolescents have about pain relief because they can control it and they can tailor the opioid delivery to the extent of pain they have at a given time, for example, before physical therapy, removal of tubes or drains, dressing changes, or getting out of bed. Additionally, the use of PCA avoids delays in administration of analgesics associated with standard “as needed” orders of IV opioids and allows smaller doses of opioids to be delivered more frequently without increasing nursing workload. Therefore, PCA is thought to provide more consistent pain relief with less total opioid dosing, resulting in fewer side effects, such as sedation, nausea, and vomiting. Purportedly, children using PCA report better analgesia and reduced pain scores compared with children who have to rely on the nursing staff to administer analgesics when they are in pain. These and other benefits of PCA have been extensively touted in the medical literature,^274–276 as well as in the lay press.²⁷⁷ Recently, however, risks associated with PCA use have also been highlighted and are discussed later.^278–280 Recognition of these risks has led to recommendations for careful dosing and monitoring of all children who are receiving opioids, particularly those receiving continuous infusions and those with specific risk factors.²⁸¹

Child training is a necessary part of PCA, because successful use of PCA requires that both the child and family understand how it works.²⁸² The instructions should be clear that the pump should be activated whenever the child feels pain, that children cannot give themselves “too much medication” because of the computer lockout interval, that the child should not wait for severe pain to activate the pump, and that a dose can also be given in anticipation of painful stimuli, such as ambulation or chest physiotherapy. Most importantly, PCA does not mean parent-controlled analgesia, and parents should never activate the pump unless specifically authorized to do so by the primary care or pain service physician (see Nurse/Caregiver-Controlled Analgesia, later).²⁸¹

Pump Settings

Most PCA pumps have five settings to adjust:

A loading dose of opioid ranging from 0.025 to 0.1 mg/kg morphine divided into incremental doses is usually given to establish adequate analgesia before therapy is turned over to the child, because self-administered doses with this technique are generally small. A sufficient interval between incremental doses must be allowed, so that the morphine achieves its peak effect before the next dose, thereby avoiding an overdose. If PCA is started in the PACU, opioid doses administered during surgery must be considered before prescribing a loading dose. Additionally, it may be desirable to administer the loading dose via the PCA pump so that it is included in the initial 4-hour or hourly limit of the PCA. That is because children who receive IV-PRN doses of opioids in the PACU, followed by initiation of PCA, may be at risk for oversedation and respiratory depression resulting from opioid stacking. Children who have received opioids toward the end of surgery, those who awaken in comfort, or those who receive nerve blocks may not need a loading dose and may start to use the demand doses as needed on awakening.

A patient bolus dose, that is, the dose that will be administered with each child’s activation of the pump, must be prescribed. These small boluses are usually in the range of 0.01 to 0.02 mg/kg of morphine in opioid-naive subjects.

A lockout interval of usually 5 to 15 minutes prevents a child from activating the pump until the full effect from the previous bolus is achieved, and it should correspond to the time from IV injection to the peak effect of the drug.

A continuous basal infusion ranging from 0.00 to 0.02 mg/kg/hr of morphine (or more, in opioid-tolerant subjects) may be used in selective cases (see later).

A maximum hourly dose or a 4-hour limit may be chosen to limit the cumulative amount of drug a child can administer. Once this limit is reached, the child cannot activate the pump until the 4-hour limit has passed. Four-hour limits allow for increased flexibility in dosing over greater periods of time and pain intensity. Typically, the maximum hourly dose ranges from 0.05 to 0.1 mg/kg and 4-hour limits from 0.25 to 0.4 mg/kg of morphine in opioid-naive subjects. This amount may be chosen based on the average hourly use of morphine during the past 24 hours or, in children started on PCA immediately after surgery, at the reduced range of the dosage scale. Figure 43-8 presents sample PCA orders, including choice of drugs, dosing, and suggested monitoring.

FIGURE 43-8 Sample patient-controlled analgesia orders.

(Modified from the University of Michigan Hospitals & Health Centers.)

Continuous Basal Infusions (CBI)

The use of a CBI of the opioid to supplement child-administered doses remains a subject of controversy. The rationale for the use of CBI is to maintain near-therapeutic plasma opioid concentrations, particularly during periods of sleep when there may be no self-administered doses, as illustrated in Figure 43-9, A. On the other hand, as depicted in Figure 43-9, B, a child who receives only PCA bolus dosing with no CBI is likely to awaken with unrelieved pain that may require multiple doses to again achieve adequate pain relief. Decreased nocturnal awakenings secondary to pain, improved restfulness or sleep patterns, reduced total opioid consumption, fewer adverse effects, and improved analgesic effectiveness have all been proposed as potential reasons in favor of using CBI. However, the use of CBI commits the child to receiving a fixed dose of opioid regardless of the level of sedation, and has the theoretical potential for overriding one of the inherent safety features of PCA, that is, an excessively sedated or somnolent child is unlikely to push the button and therefore receives no additional opioid but, with a fixed infusion, drug may accumulate (Fig. 43-9, C ), with the potential for hypoventilation.²⁸⁶ Furthermore, it has also been argued that programming errors with CBI can lead to more serious adverse events because the opioid medication is delivered regardless of the child’s level of sedation.^285,287

FIGURE 43-9 Patient-controlled analgesia (PCA) with an appropriate dose of continuous basal infusion. A, Use of an appropriate dose of continuous basal infusion (CBI) maintains near-therapeutic plasma opioid concentrations during sleep with rapid increase to analgesic concentrations as soon as the child awakens and activates the PCA button. B, Chart depicts plasma opioid concentrations in a child who is receiving PCA bolus doses only, with no CBI. Plasma concentrations decrease significantly when the child is asleep, necessitating multiple doses to achieve analgesic concentrations. C, An excessive dose of CBI results in opioid accumulation with delayed hypoventilation even when the PCA button is not activated.

(Reproduced with permission from Berde CB, Solodiuk J. Multidisciplinary programs for management of acute and chronic pain in children. In: Schecter NL, Berde CB, Yaster M, editors. Pain in infants, children and adolescents. 2nd ed. Philadelphia: Lippincott Williams and Wilkins; 2003, p. 476.)

Some adult studies have suggested that the use of CBI has limited benefit in terms of efficacy and is associated with a greater incidence of opioid adverse effects, including respiratory depression.^288–290 Studies in children, however, have yielded conflicting results.^{273,274,291–295} Children 7 to 19 years of age who received PCA with CBI after orthopedic surgery reported significantly reduced pain scores compared with those who received PCA boluses alone or IM morphine.²⁷⁴ There were no differences in morphine consumption or in opioid adverse effects among the three groups, with no incidents of respiratory depression. Notably, child satisfaction was greatest in the PCA with CBI group. Similar pain scores with improved sleeping patterns have also been reported with the use of PCA with CBI, compared with those who received PCA alone, on the first two postoperative nights in children after abdominal surgery. No incidents of respiratory depression or excessive sedation were reported in either group.²⁷³ Children who received CBI with PCA or NCA in one study reported slightly reduced pain scores without differences in morphine use or adverse effects after spine fusion surgery compared with those who received PCA or NCA alone.²⁹⁵ In contrast, others have reported greater morphine use and a greater incidence of hypoxemia with similar pain scores in children who received PCA plus CBI after surgery compared with those who received PCA alone.^293,294 A subsequent study found that children who received PCA with a CBI of morphine (4 µg/kg/hr) experienced fewer adverse effects and less hypoxemia than children who received PCA with 20 µg/kg/hr CBI or those who received PCA bolus doses alone, noting similar pain scores in all three groups.²⁹² Both of the PCA with CBI groups reported better sleep at night than did the PCA alone group. Differences in opioid doses, age groups, and surgical procedures may account, in part, for these apparent conflicting observations in the pediatric studies. Based on these studies and our own experience, we hold the view that the use of CBI may be beneficial in some children, although it requires careful selection of dose, based on the surgical severity and the child’s comorbid conditions and vigilant monitoring to ensure safety. CBI should be used as a routine for most children with pain resulting from cancer, for most children with mucositis resulting from cancer treatment or bone marrow transplantation, and for a significant percentage of children with sickle cell vasoocclusive episodes. At our institutions, the standard practice is to use CBI, unless limited by somnolence or hypoventilation, for the first night for a majority of children undergoing selected major painful surgeries, such as scoliosis surgery, pelvic osteotomies, and thoracotomies.

Nurse/Caregiver-Controlled Analgesia

Activation of the PCA pump by the bedside nurse, a parent, or a caregiver (such as a grandparent) has been used with success in children who are unable to push the button because of young age or because of physical or cognitive impairments.^{275,276,295–297} In much of the literature, and in some policy statements by the Joint Commission and other organizations, there is, in our view, too little distinction between activation of PCA pumps by nurses and activation by nonclinician surrogates. The term caregiver is used variably in this literature; here it is used to mean nonclinician surrogates. In pediatrics, this means primarily parents and other family members. A small study of 12 children who received NCA after spine fusion surgery reported adequate analgesia, parent and nurse satisfaction, and no complications.²⁹⁵ Children who received NCA received smaller total morphine doses compared with those who were able to use the PCA device themselves, likely because of the tendency for nurses to underestimate their patients’ pain. A larger observational study of 212 children who received parent- or nurse-controlled analgesia with morphine, fentanyl, or hydromorphone reported effective analgesia (pain scores 3/10 or 2/5 or below) in more than 80% of children.²⁷⁶ Pruritus occurred in 8% and vomiting in 15% of the children on the first day of treatment. Nine children (4.2%) required naloxone for the following: apnea (N = 4), hypoxemia (N = 1), excessive sedation (N = 3), or to facilitate extubation (N = 1). Six of these children had significant comorbid conditions, and 5 received additional sedatives. These investigators emphasized the importance of close monitoring to minimize risk and permit early intervention when using nurse/caregiver-controlled analgesia (NCA/CCA). Another study found that the incidence of overall adverse events in opioid-naive children who received NCA was similar to that in children who received PCA after surgery (22% and 24%, respectively).²⁹⁷ However, children who were able to self-administer PCA required only minor interventions (stimulation, reduction in opioid dosage, or supplemental oxygen), whereas those who received NCA were more likely to require more aggressive interventions, such as opioid reversal, airway management, or escalation of care. This study found that cognitive impairment and opioid dose on the first postoperative day were independent predictors of adverse events. Although the mean time to the occurrence of adverse events was 16 to 27 hours, some events occurred during the third postoperative day, suggesting that monitoring, including continuous pulse oximetry, should be continued as long as the PCA is used. In a study of children with cancer, five respiratory and/or neurologic serious adverse effects were reported, with 1 child requiring naloxone during the 576 days of treatment with NCA/CCA.²⁹⁶ In that study, pulse oximetry was used only at the discretion of the provider, perhaps reducing the investigator’s ability to recognize hypoxemia in some children. Furthermore, the reduced incidence of adverse events may be explained by the fact that most children in that study were not opioid naive and may have developed some degree of opioid tolerance. A study of 10,000 pediatric patients receiving NCA further confirms the safety and tolerability of this delivery method.¹⁰⁹

For safety reasons, however, it is important to distinguish between authorized and unauthorized use of the PCA button by an individual other than the child. Several reports describe serious adverse events, including excessive sedation, severe respiratory depression, respiratory arrest, and death, attributed to unauthorized activation of the PCA device by parents, spouses, other family members, and health care providers.^297–301 The practice of PCA by proxy has therefore come under scrutiny and its safety questioned by the Joint Commission and the Institute for Safe Medication Practices (ISMP).^302–304 In 2004, the Joint Commission issued a sentinel event alert based on PCA errors reported to the U.S. Pharmacopeia. Of 460 errors that resulted in death or some level of harm to the patient, 15 resulted from PCA by proxy, including 12 attributed to family members, 2 to a nurse, and 1 to a pharmacist.³⁰² In interpreting this report and some other reports, it should be emphasized that many do not cite denominator data (total numbers of patients receiving PCA vs. NCA/CCA), so that it is difficult to assign numbers for relative or absolute risks of these techniques. Recognition of these risks has led the Joint Commission and ISMP to strongly recommend that specific policies and procedures be developed and implemented related to the use of PCA by individuals other than the child. Such policies must address the following issues:

With carefully defined policies and procedures, adequate education of clinicians and caregivers, prevention of unauthorized dosing, and vigilant monitoring, it may be possible to reduce the frequency of adverse events from PCA, NCA, and, especially, CCA. However, large outcomes studies are needed after implementation of such policies to confirm the safety and efficacy of practice based on these recommendations. Our view is that NCA is a well-established practice and should be encouraged as a generally safe and effective means for delivering opioids to children who are unable to self-administer.

Risks and Adverse Events with PCA

Despite its numerous benefits, the use of PCA has been associated with a wide range of adverse effects, adverse events, and unfavorable outcomes in adults.^{278,279,299,305–308} Although some adverse effects from PCA therapy may be attributed to the opioid drugs themselves or to patient comorbidities, a significant number of harmful effects occur as a result of human error, with incorrect prescribing, dispensing, administration, or equipment failure. E-Table 43-5 describes the causes of PCA medication errors as identified by the ISMP.^303–305309 Increasing awareness of preventable adverse events from PCA has led to improved pump technology directed at minimizing the likelihood of programming errors, including the development of smart PCA pumps that use bar-coded syringes and an integral bar-code reader to prevent incorrect programming of drug concentration. Potential PCA pump errors have been reduced with these “smart pumps.”³¹⁰ Children are at greater risk for medication-related adverse events resulting from calculation errors in drug doses (because all doses are based on body weight or body surface area) and to developmental differences in PK. However, data for PCA-related adverse events in children are limited.^{280,311–313} The reported incidence of respiratory depression in children receiving PCA ranges from 0% to 25%.^{297,311,312,314,315} Risk factors for respiratory depression identified by these studies include cumulative opioid dose, use of basal infusions, concomitant administration of sedatives, and comorbid conditions, including renal failure and cognitive impairment. Recognition of the risks from PCA^316,317 has led organizations, such as the ISMP, to emphasize the importance of monitoring children who use PCA and of detailed child and staff education regarding its use.³¹⁸

E-TABLE 43-5 Causes of Patient-Controlled Analgesia Errors

Single-Injection Techniques

There are many circumstances in which the simplicity and duration of action of a single-injection block is desirable. Both neuraxial blocks and peripheral plexus and nerve blocks can be effective as single-injection techniques. The caudal block remains the most commonly used technique in pediatric regional anesthesia. It provides analgesia of the lower extremity and lower abdomen, and is easily and quickly performed in most infants and children (see Chapter 41). When an operation is performed on both legs it is often preferable to bilateral lower extremity blocks.

The use of ultrasound makes the identification of plexuses and peripheral nerves much easier in an anesthetized child. Peripheral nerve stimulation using Teflon-coated needles remains an effective technique, but is becoming less frequently used as familiarity with ultrasound and its benefits increases. It is plausible, but unproven, that ultrasound reduces the risk of injury to nerves and adjacent structures (see Chapters 41 and 42).⁴⁰⁴ To precisely position the needle in the target area and avoid piercing adjacent structures, one must develop skill in coordinating its trajectory with the ultrasound image so that the needle tip does not pass out of the plane of view. Although it remains unknown whether ultrasound and more precise needle placement will reduce nerve injury, there is evidence that it reduces the volume of local anesthetic needed to produce an effective block.^405,406 Ultrasound visualization of the local anesthetic surrounding a nerve or plexus provides reliable confirmation that a block will be successful. Nerve blocks of the lower extremity can, in some circumstances, be used instead of caudal blockade (see Chapters 41 and 42). They provide a field of analgesia limited to the operative site, have a prolonged duration of analgesia (usually at least twice that of caudal block), and eliminate some of the potential undesirable effects of central neuraxis blockade, such as urinary retention, motor blockade (wobbly legs), and, occasionally, numbness. The fascia iliaca block, which produces analgesia in the distribution of the femoral, obturator, and lateral femoral cutaneous nerves, has been described in children as an excellent alternative for postoperative analgesia of a lower extremity.⁴⁰⁷ Blockade of the femoral nerve, the sciatic nerve, and the nerves of the ankle is easily performed in children by using similar techniques to those in adults (see Chapters 41 and 42). Regional blockade of the upper extremity may involve an axillary block, interscalene, supraclavicular, or infraclavicular block, or less commonly, blockade of the individual nerves at the level of the arm or wrist. Intercostal nerve blocks may be used after thoracic or upper abdominal surgery. This block may be considered for procedures of limited scope, such as open-lung biopsy or thoracostomy for drainage, but the duration of blockade is limited to several hours. In our view, for most thoracic procedures, the duration of single-injection percutaneous intercostal blockade is generally too short to warrant the risk of pneumothorax. For open-chest procedures, intercostal blocks by the surgeon pose a low risk and may be indicated, although again, their effectiveness is limited by their short duration. Epinephrine is commonly included in the local anesthetic for these blocks to limit the rate of uptake. More extensive surgery anticipated to cause postoperative pain of longer duration may be best managed with a catheter (continuous infusion) technique.

Ilioinguinal-iliohypogastric nerve block provides excellent postoperative analgesia for inguinal herniorrhaphy, a common outpatient procedure in children (see Chapters 41 and 42). It appears similar in efficacy to a caudal block, with a duration of analgesia of at least 4 hours when bupivacaine with epinephrine is used. For orchiopexy, ilioinguinal nerve block has been found to be as effective as caudal blockade to the T10 level in a randomized and blinded investigation.⁴⁰⁸ In our experience, however, postoperative analgesia for procedures that involve considerable manipulation and traction on the spermatic cord and testis may be better managed with caudal blockade. Penile block is effective for both circumcision and distal, simple hypospadias repair. More extensive procedures on the penis, especially repair of penile-scrotal hypospadias, require a caudal, rather than a penile block, to produce effective analgesia.⁴⁰⁹ The use of single-injection neuraxial opioids is an additional technique that can provide longer-lasting analgesia, but must be used in inpatients only, because of the need for respiratory monitoring. Intrathecal morphine has been administered to infants and children for several decades and provides long-acting analgesia (up to 24 hours) after a single injection.^410,411 It is absolutely essential that only preservative-free preparations of the drug be used because preservative-containing solutions may result in injury to the central neuraxis. Doses ranging from 5 to 10 µg/kg are usually chosen. Because of the hydrophilicity of this agent, respiratory monitoring is mandated. Respiratory depression, as well as pruritus, nausea, and urinary retention, are reported after intrathecal morphine. We most commonly employ this modality for analgesia after posterior spinal fusion. The drug is often administered by the surgeon under direct vision of the dura, but if administered at the beginning of the case, it has been reported to reduce intraoperative blood loss.^412,413 This route of administration has also been employed before cardiac surgery.⁴¹⁴ The effective intrathecal dose of opioid is roughly one-fifth to one-tenth that of the epidural dose, and the duration of action, especially with morphine, is significantly prolonged. Therefore, observation in a monitored setting must be continued for 24 hours or until no further evidence of respiratory compromise exists (without the use of naloxone). A novel extended-release preparation of morphine uses a liposomal foam suspension of the drug (DepoDur, EKR Therapeutics, Bedminster, N.J.) that is injected in the lumbar epidural space (note that it should not be used intrathecally) and provides 48 hours of analgesia, with no increase in untoward effects compared with conventional neuraxial morphine.^415–418 To date there are no pediatric data, but clinical trials in children are currently in progress.

Catheter Techniques

A catheter placed in the epidural space or adjacent to a nerve or plexus can be used to provide continuous uninterrupted analgesia for prolonged periods after surgery. These catheters are commonly used for about 3 days but may be used for more prolonged periods in selected situations. With proper care, catheters tunneled under the skin may be left in situ for more than 7 days.^419,420 Catheters can also be placed in intrapleural or extrapleural and/or retropleural locations to provide analgesia after thoracic surgery.⁴²¹ In our experience, intrapleural analgesia reduces but does not eliminate the need for systemic opioids, especially if thoracostomy drains are present. We rarely use this technique because toxic concentrations of local anesthetic and seizures have been reported.⁴²² Continuous extrapleural, intercostal, and paravertebral catheters can be placed either through the operative field or percutaneously. In adult studies, and in a smaller series of pediatric studies, these techniques appear to provide excellent analgesia at rest, but with movement they provide only partial opioid sparing. Some clinicians regard these techniques as providing many of the advantages of thoracic epidural analgesia, with the potential for reducing the risks and adverse effects of thoracic epidural analgesia.

Continuous regional blockade is remarkably effective and safe, although as with any technique, monitoring for untoward effects is necessary to prevent complications. New technology for the delivery of drugs to peripheral nerves and plexuses have made it possible for children to receive the benefits of continuous neural blockade after discharge from the hospital or day surgery unit.³³³ Catheters may deliver local anesthetics and other medications via controlled infusion devices that use a pressurized elastomer-bulb reservoir that controls the infusion rate with a flow limiter (ON-Q pump, I-Flow LLC, Lake Forest, Calif.; Infusor, Baxter, Deerfield, Ill.; Accufusor, Moog Medical Devices, Salt Lake City; Easypump, B Braun Melsungen AG, Melsungen, Germany; and others) (E-Fig. 43-2). These devices can be used at home after discharge, for infusion of medication into a tissue plane to provide a continuous field blockade or a continuous peripheral nerve block. Infusion of local anesthetics in the subcutaneous tissues at the incision site or into the surgical plane using these devices provides prolonged analgesia in both adults and children. Several types of these infusion systems are available, including ones that have options for fixed, variable, or continuous infusion rates with a bolus option. The latter two must be used with caution in smaller children in whom local anesthetic toxicity from excessive dosage may be a risk; we generally use the fixed-rate devices. When continuous peripheral nerve or plexus blocks are used for outpatients, a carefully designed system must be in place to follow these children, to avoid complications and achieve early detection of potentially adverse events. Daily follow-up, either by visiting nurses or by phone, is mandatory. We are aware of a small number of anecdotal cases of possible excessive doses of local anesthetic from a continuous infusion pump used at home that resulted in adverse effects, although no catastrophic events have been reported.

E-FIGURE 43-2 ON-Q pump (I-Flow LLC, Lake Forest, Calif.): an elastomeric infusion device for ambulatory continuous regional blockade. The pump bulb is filled with local anesthetic and pressurized. The flow restrictor is readily set to the desired flow rate, and the system is disposable.

Caudal and Epidural Catheters.

With experience and proper equipment, the lumbar route is feasible at any age, but specific expertise is required for infants and toddlers. Practitioners who have less experience with lumbar epidural catheterization in children should consider placing the catheter via the caudal route for children less than 6 years of age. In infants and children up to about 6 years of age, catheters may be advanced freely from the caudal canal cephalad to lower thoracic levels with excellent success.⁴²³ This is possible in part because young children have a less developed vascular plexus and more compact and globular fat than do older children and adults.^423,424 However, other authors describe less reliability with caudal-to-thoracic advancement of catheters in children larger than 10 kg.⁴²⁵ This has been attributed to the more mature composition of the contents of the epidural space and the lumbar lordosis that occurs with walking. With age, the epidural fat appears to lose the spongy gelatinous character noted in infants and the spaces between the fat globules become less distinct.^423,424 Catheters made of nylon or polyamide may be less likely to kink beneath the skin and seem to thread more easily than those made of Teflon or other materials. The catheter should never be advanced if resistance is felt. Similar difficulties in advancing catheters that resulted in catheters looping backwards, or kinking, or puncturing the dura have been reported in neonates weighing less than 3.5 kg.⁴²⁶ In one series of 20 preterm infants, epidurography revealed misplaced catheters in 3 infants or 15%. New data suggest that misplacement of threaded catheters in all ages may be more common than is generally recognized.⁴²⁷ In this series of 724 epidurograms, unexpected misplacement was detected in 11 or 1.5%, including 3 intravascular catheters, despite negative test doses, 2 intrathecal without cerebrospinal fluid aspiration, 4 that were intraperitoneal, and 1 each in the rectum and psoas compartment. These authors recommended epidurography in all cases, although this is not the current standard of practice. If specific dermatomal placement of a catheter is sought, one should consider obtaining an imaging study to confirm the dermatomal level of the tip of the catheter, and the aforementioned data suggest that imaging might be advisable whenever a catheter is advanced to a level substantially more rostral than its insertion level.

As an alternative to epidurography, a nerve-stimulating catheter may be used, which allows real-time monitoring of the location of the tip of the catheter as it is advanced cephalad (with the stylet in place, using one system available in Canada), or at least confirmation of tip location on removal of the stylet (using a convenient modification of equipment readily available in the United States).^428,429 The technique requires the use of saline loss-of-resistance and avoiding air bubbles in the epidural space or in the catheter-connector-injection system, because air impedes electrical conduction. In addition, neuromuscular blockade must be avoided because it abolishes the motor response. Used in conjunction with a nerve stimulator set to very low milliamperage (approximately 6 mA), the muscles supplied by a nerve root will twitch as the catheter approaches the segments that supply that dermatome, thereby confirming the catheter tip’s location. Specifically, twitches in the feet and ankles occur with catheter tips around L5 to S1, hip flexion occurs with catheter tips around T12 to S1, abdominal muscle twitches without hip flexion imply thoracic positioning above T12, intercostal muscle twitches imply midthoracic tip positioning. Finger twitches would imply advancement to around T1. This technique can also be used to detect catheter malpositioning. Bilateral twitching at a current less than 0.6 mA generally implies subarachnoid positioning. Unilateral twitches in a narrow motor distribution at a current less than 1 mA may suggest advancement out a root foramen. Unilateral twitches at a current less than 1 mA in a very broad motor distribution may indicate subdural positioning. Absence of twitches as the current is increased to about 15 mA (in the absence of air bubbles or neuromuscular blockade) generally indicates that the epidural catheter is not in the epidural space. Our experience is that the use of one of these confirmatory techniques can help avoid problems with failed or incomplete blocks in the postoperative period.

Choice of drugs for epidural infusions depends on a number of factors, including site of surgery, site of the epidural catheter tip, and child risk factors. Local anesthetics, lipophilic opioids, and, to some extent, clonidine all have more restricted cephalad distribution of action during infusions, compared with hydrophilic opioids, such as hydromorphone or morphine. Consequently, optimal positioning of the catheter tip during insertion can improve the analgesic action postoperatively. Experience in adults suggests that when local anesthetics are used, analgesia is optimized when the tip of the epidural catheter is positioned at or slightly above the dermatomal levels involved in the surgery. This effect may be more pronounced when children are moving than when they are at rest. Positioning at levels slightly above the dermatomal levels involved in surgery is more relevant for surgery in lumbosacral dermatomes compared with upper thoracic dermatomes because of the greater distance between root level and dorsal horn level in these two circumstances.⁴³⁰

Placement of catheters at thoracic levels requires consideration of risk-benefit trade-offs. Therefore, if a catheter tip is at the lumbar or caudal level and the surgery is in thoracic and upper abdominal dermatomes, continuous infusions containing local anesthetics alone or with lipophilic opioids, such as fentanyl or sufentanil, are likely to be ineffective. Increasing the infusion volume is limited by maximum allowable systemic local anesthetic concentrations. In this circumstance, a reasonable alternative is to administer hydrophilic opioids (e.g., morphine or hydromorphone) through lumbar or caudal catheters. The hydrophilic properties of these opioids permit a greater rostral spread, such that a wider range of dermatomes can be covered after caudal or lumbar catheter placement, with only small increases in dose for surgical sites remote from the catheter tip. This same characteristic, unfortunately, also appears to increase the risk for adverse effects, including respiratory depression, as a result of rostral spread of morphine to the central respiratory center in the brainstem. Hydromorphone may offer some advantages because its rostral spread is intermediate compared with fentanyl and morphine and it causes slightly less pruritus compared with morphine.^431,432 Lumbar administration of hydromorphone spreads sufficiently to provide thoracic analgesia.⁴³³ It should be noted that the potency of hydromorphone relative to morphine in the epidural space, is less than when it is administered systemically (2-3 : 1 epidural vs. 5 : 1 systemic).⁴³⁴ Such a technique, however, precludes the optimal use of local anesthetics, and the considerable benefit that may accrue from a multimodal approach to regional analgesia.

A thoracic epidural catheter can be placed in an awake, sedated, or anesthetized child. Some clinicians are hesitant to perform thoracic epidural puncture in anesthetized children because of the inability to use child reports (e.g., paresthesia or lancinating pain) as an indicator of improper catheter or needle placement. However, there are few, if any, data to suggest that the risk of inserting a thoracic epidural catheter in an anesthetized child by an experienced clinician is actually greater than attempting such placement in an awake child who may not be able to remain still, nor is it clear that a well-sedated child can accurately report sensations of paresthesia or differentiate the discomfort of needle passage from more ominous pain.⁴³⁵ The prospective data from the British National Epidural Audit do not support the contention that there is increased risk with this practice, and those data are corroborated by the PRAN study, which found no neural injury regardless of patient state during epidural placement.^436–438 We believe that data and experience support the safety of placement of thoracic epidural catheters in anesthetized children by anesthesiologists trained and experienced in this procedure, although the tolerances for error are smaller than in adults and older children.^439–441

It is extremely important to secure caudal catheters to the skin with a clear occlusive dressing, to avoid contamination or dislodgment and to allow daily inspection of the site. In addition, tincture of benzoin or other adhesive solution reduces the incidence of catheter and dressing displacement. For a caudal catheter, the use of an adhesive-edged plastic drape, covering the area from the gluteal crease over the dressing, also helps prevent fecal soiling of the dressing by children who are in diapers. Despite these precautions, if a dressing becomes detached and the insertion site is contaminated it is prudent to remove the catheter. The use of lumbar catheters removes the insertion site from the diaper area and thereby further reduces the potential for contamination of the catheter and the insertion site.

Successful management of children with an epidural catheter requires carefully coordinated monitoring protocols, nursing management, and medical management of drug selection and dosing. As with PCA or continuous-infusion opioids, the orders should be standardized and written in consultation with the nursing staff, so that misinterpretations are less likely to occur (Fig. 43-10). Because the single most sensitive monitor of children receiving epidural opioids is the nurse, rather than a mechanical or electronic device, education of the nursing staff is of paramount importance to ensure safety. Nursing staff can also assess the adequacy of analgesia and thereby help to titrate the drug dose. Catheter insertion sites should be inspected at least once daily by the pain service, both for the integrity of the dressing and for any evidence of erythema or skin infection (Fig. 43-11). When continuous infusions are used, the tubing connecting the infusion pump to the catheter should not have any injection ports and it should be clearly labeled as an epidural catheter to preclude unintended epidural administration of drugs intended for IV use. We use color-coded tubing and a color-coded drug cassette for the continuous administration of regional agents. Although the same pumps are used for IV PCA administration, it is immediately obvious what route the drug is intended for simply by recognizing the tubing and cassette color.

FIGURE 43-10 Sample epidural orders.

(Modified from the University of Michigan Hospitals & Health Centers.)

FIGURE 43-11 Superficial skin infection at the site of an epidural catheter insertion. The catheter was withdrawn, and the problem resolved with local skin care only.

A variation of traditional epidural analgesia and PCA is the technique of patient-controlled epidural analgesia (PCEA).⁴⁴² With this variation of traditional PCA, children are generally maintained with an infusion of epidural analgesics (frequently opioid alone or an opioid–local anesthetic mixture) and have the capability of self-administering supplemental doses when needed. When using this technique, the background infusion is used to provide the majority of the analgesia and the child can add to this when needed. It must be emphasized that the time needed for a bolus dose to effect a change with epidural administration is more prolonged than it is with IV agents. Therefore, with PCEA, lockout intervals are greater (often 15 to 30 minutes) than with PCA.^443,444 The considerations one would employ for choosing this technique include the same child-monitoring factors for IV PCA and epidural analgesia, and maximum local anesthetic doses must be carefully calculated to cover the contingency of the greatest possible activation of the PCEA demand doses. “Smart pump” technology has been demonstrated to improve safety of this technique.⁴⁴⁵