Total body water and extracellular fluid (ECF) [29] decrease throughout gestation, the neonatal period, and childhood (Fig. 3.3), whereas the percent of body weight contributed by fat increases from 3 % in a 1.5-kg preterm neonate to 12 % in a full-term neonate and then doubles by 4–5 months of postnatal age. These changes in the body composition substantively affect the volumes of distribution of drugs. Polar drugs such as depolarizing and non-depolarizing neuromuscular-blocking drugs (NMBDs) distribute rapidly into the ECF, but enter cells more slowly. The initial dose of such drugs is consequently greater in the neonate than in the child or adult. Lipid-soluble drugs may also have a larger volume of distribution in neonates. The volume of distribution at steady state of fentanyl is 5.9 (SD 1.5) L/kg in a neonate compared with 1.6 (SD 0.3) L/kg in an adult [30]. This may explain the reduced respiratory depression after doses as large as 10 μg/kg in full-term neonates. However, high-dose therapy (50 μg/kg) results in a prolonged effect in neonates due to reduced clearance. In the case of propofol, reduction in the plasma concentration after induction of anesthesia is attributable to redistribution rather than rapid clearance. Neonates have less body fat and muscle content than older children. Hence, less propofol is apportioned to these “deep” compartments, attenuating the redistribution of propofol. If repeat doses of propofol are administered, they may accumulate in the blood and brain and delay recovery.

The plasma alpha 1-acid glycoprotein (AAG) and albumin concentrations in neonates are reduced, albeit with a broad range of scatter (e.g., AAG 0.32–0.92 g/L), but reach adult concentrations by 6 months of postnatal age [31–33]. AAG is an acute phase reactant that increases after surgical stress. This increases the total plasma concentrations for low to intermediate extraction drugs such as bupivacaine that bind to AAG [34].

The concentration of unbound bupivacaine, however, will not change because the clearance of unbound bupivacaine depends only on the intrinsic metabolizing capacity of the liver. Any increase in unbound concentrations, for example, during long-term epidural infusions, is attributable to a reduced clearance rather than a decrease in the AAG concentration [35]. Total bupivacaine concentrations increase in the first 24 h after surgery in neonates who are receiving a continuous epidural infusion of bupivacaine; unbound bupivacaine concentrations, however, have not been studied. This increase in total bupivacaine is attributable in part to an increase of AAG. This increase in total bupivacaine, combined with reports of seizures in infants given epidural bupivacaine infusion, has led to recommendations to stop epidural infusions at 24 h. However, it is the unbound bupivacaine concentration that is responsible for the CNS effects, and this is determined by the clearance of the unbound bupivacaine. Clearance, the pivotal variable in the elimination of unbound bupivacaine, is reduced in neonates. In addition, clearance shows large interindividual variability, which means that unbound bupivacaine concentrations may increase steadily in some individuals with very low clearance. The lack of knowledge regarding the clearance of bupivacaine in each neonate precludes pronouncing the duration of a safe bupivacaine infusion for all [36].

Plasma albumin concentrations are least in preterm neonates but increase steadily, approximating adult values by 5 months of postnatal age. Binding capacity approaches adult values by 1 year of age. In addition, free fatty acids and unconjugated bilirubin compete with acidic drugs (e.g., ibuprofen, ceftriaxone) for albumin binding. The induction dose of thiopental in neonates is less than it is in children. It is possible that this is related to the decreased binding of thiopental to plasma albumin; 13 % of the drug is unbound in neonates compared with 7 % in adults [37].

The initial phase of distribution reflects the magnitude of regional blood flow. Consequently, the brain, heart, and liver, which receive the largest fraction of cardiac output, are first exposed. Drugs are then redistributed to other relatively well-perfused tissues, such as the skeletal muscle. There is a much slower tertiary distribution to relatively under-perfused tissues of the body that is noted with long-term drug infusions. In addition to perinatal circulatory changes (e.g., ductus venosus, ductus arteriosus), there are maturational differences in relative organ mass and regional blood flow, while a symptomatic patent ductus arteriosus may also result in differences in distribution. Blood flow as a fraction of the cardiac output, to the kidney and brain, increases with age, whereas blood flow to the liver decreases through the neonatal period [38]. Cerebral and hepatic mass as proportions of body weight in the infant are much greater than in the adult [39]. While onset times are generally faster for neonates than adults (a size effect), reduced cardiac output and cerebral perfusion in neonates mean that the expected onset time after an intravenous induction is slower in neonates, although reduced protein binding may counter this observation for some drugs. Offset time is also delayed because redistribution to well-perfused and deep under-perfused tissues is more limited.

The blood-brain barrier (BBB) is a network of tight junctions that restricts paracellular diffusion of compounds between the blood and brain. Confusion over the importance of this barrier in the neonate exists, partly because of early studies on respiratory depression caused by morphine and meperidine [40]. Early investigations found that the respiratory depression after morphine was greater than that after meperidine. This difference was attributed to greater brain concentrations of morphine because of the poorly developed BBB in the neonate [40]. It was postulated that BBB permeability to water-soluble drugs, such as morphine, changes with maturation [40]. However, the neonatal respiratory depression observed after morphine could have been explained by pharmacokinetic age-related changes. For example, the volume of distribution of morphine in term neonates 1–4 days (1.3 L/kg) is reduced compared with that in infants 8–60 days of age (1.8 L/kg) and in adults (2.8 L/kg) [41]. Consequently, we might expect greater initial concentrations of morphine in neonates than in adults, resulting in more pronounced respiratory depression in the former. Respiratory depression, measured by carbon dioxide response curves or by arterial oxygen tension, is similar from 2 to 570 days of age at the same morphine blood concentration [42]. The BBB theory in this particular circumstance lacks strong evidence. It is more likely that the increased neonatal respiratory depression after morphine is due to pharmacokinetic age-related changes.

The BBB may have impact, however, in other ways. Small molecules are thought to access fetal and neonatal brains more readily than in adults [43]. BBB function improves gradually, possibly reaching maturity by full-term age [43]. Kernicterus, for example, is more common in preterm neonates than in full-term neonates. In contrast to drugs bound to plasma proteins, unbound lipophilic drugs passively diffuse across the BBB equilibrating very quickly. This may contribute to bupivacaine’s propensity for seizures in neonates. Decreased protein binding, as in the neonate, results in a greater proportion of unbound drug that is available for passive diffusion.

In addition to passive diffusion, there are specific transport systems that mediate active transport. Pathological CNS conditions can cause BBB breakdown and alter these transport systems. Fentanyl is actively transported across the BBB by a saturable ATP-dependent process, whereas ATP-binding cassette proteins such as P-glycoprotein actively pump out opioids such as fentanyl and morphine [44]. P-Glycoprotein modulation significantly influences opioid brain distribution and onset time, magnitude and duration of analgesic response [45]. Modulation may occur during disease processes, fever, or in the presence of other drugs (e.g., verapamil, magnesium) [44]. Recent evidence identified P-glycoprotein in the brain of fetuses as early as 22 weeks’ gestation [46]. The prevalence and concentration of P-glycoprotein increased with fetal maturation. Genetic polymorphisms that affect P-glycoprotein-related genes may explain differences in CNS-active drug sensitivity [47, 48].

Constitutional, environmental and genetic factors all contribute to the variability in clearance, but in the young neonate, age is the dominant covariate. Most CYP isoenzymes have small phenotypic activity until birth except for CYP3A7 [49, 50]. CYP3A7 expresses its greatest activity during gestation after which its activity steadily decreases until there is no activity by 2 years [51, 52]. CYP2E1 activity surges after birth [53], CYP2D6 becomes detectable soon thereafter reaching adult levels by 2 weeks of age, and CYP3A4 and CYP2C (Fig. 3.4) are detectable during the first week, whereas CYP1A2 is the last to appear [52, 54]. Neonates depend on the immature CYP3A4 for levobupivacaine or midazolam clearance and on CYP1A2 for ropivacaine clearance, dictating reduced epidural infusion rates in this age group [1, 2, 55, 56].

CYP1A2 is 4–5 % in the neonate reaching 50 % of adult activity levels by 1 year of age [22]. Formation of the M1 metabolite of tramadol, a reflection of CYP2D6 activity [14], appears rapidly at term and reaches 84 % of mature values by 44 weeks’ PMA.

Pharmacogenomics (PG) investigates variations of DNA and RNA characteristics related to drug response that incorporates both PK and PD. Large interindividual PK variability depends to a large extent on polymorphisms of the genes that encode for metabolic enzymes [57, 58].

The effect of genetic variability on plasma cholinesterase activity and its effect on the termination of action of succinylcholine is a well-known example [59]. Another example is the CYP2D6 single nuclear polymorphism (SNP), inherited as an autosomal recessive trait that may result in poor analgesia from codeine because the active metabolite morphine is not formed [58, 60]. Both PMA and CYP2D6 activity score explained the interindividual variability in tramadol metabolism (Fig. 3.5) [51, 60]. The interplay between maturation of tramadol clearance, M1 metabolite formation and maturing GFR on M1 concentration (and subsequent analgesia) is shown in Fig. 3.6 [48]. This observation illustrates the potential relevance of polymorphisms in neonatal pharmacology.

Some phase II isoenyzmes are mature in full-term neonates at birth (sulfate conjugation), whereas others are not (acetylation, glycination, glucuronidation) [52, 61]. Glucuronidation is important in the metabolic clearance of drugs (paracetamol, morphine, propofol) frequently administered by anesthesiologists. In the neonate, glucuronidation accounts for only 25 % of the phase II conjugation of acetaminophen, whereas in adults it accounts for ~75 % [22]. Allometric body-size scaling complemented by maturation models [8, 62] has been used to unravel the effects of maturation of the pharmacology of morphine [63–65] and paracetamol [66, 67]. Both drugs are cleared by specific isoforms (UGT1A6 and UGT2B7) [52]. In both instances, clearance is immature in the preterm 24-week PMA neonate and matures to reach adult rates by the end of the first year of life (Fig. 3.7). Dexmedetomidine is also cleared predominantly by the UGT system (UGT1A4 and UGT2B10) and has a similar maturation profile [68].

Glucuronidation is the major metabolic pathway for propofol. This pathway is immature in neonates, although multiple CYP isoenzymes (CYP2B6, CYP2C9, CYP2A6) also contribute to its metabolism and cause a more rapid maturation profile than expected from glucuronidation alone [69] (Fig. 3.7). Urine collections after intravenous bolus of propofol in neonates (PNA 11 days, PMA 38 weeks) support this contention. Urinary metabolites included both propofol glucuronide and 1- and 4-quinol glucuronide in a ratio of 1:2. Hydroxylation to quinol metabolites was active in these neonates [70], contributing to the rapid increase in clearance at this age that is faster than that reported for glucuronide conjugation alone (e.g., paracetamol, morphine).

Disease characteristics also contribute to the variability in UGT-related clearance. Maturation of morphine clearance occurs more quickly in infants undergoing noncardiac surgery compared with those after cardiac surgery [71]. Neonates requiring extracorporeal membrane oxygenation [72] or positive pressure ventilation [64] also have reduced clearance. Similarly, the clearance of propofol is reduced after cardiac surgery in children [73].

Many drugs undergo metabolic clearance at extrahepatic sites. Remifentanil and atracurium are degraded by nonspecific esterases in tissues and erythrocytes, and these processes appear mature at birth, even in preterm neonates [74]. Clearance, expressed per kilogram, is increased in younger children [75–79], likely attributable to size because clearance is similar when scaled to a 70-kg person using allometry [75]. Ester local anesthetics are metabolized by plasma butyryl-cholinesterase, which is thought to be reduced in neonates. The in vitro plasma half-life of 2-chloroprocaine in umbilical cord blood is twice that in maternal blood [80], but there are no in vivo studies of the effects of age on its metabolism. Succinylcholine clearance is increased in neonates [81, 82], suggesting butyryl-cholinesterase activity is mature at birth.

The factors that determine anesthetic absorption through the lung (alveolar ventilation, FRC, cardiac output, solubility) also contribute to elimination kinetics. We might anticipate more rapid washout in neonates for any given duration of anesthesia because of the greater alveolar ventilation to FRC ratio, greater fraction of cardiac output perfusing vessel-rich tissues, reduced solubility in blood and tissues, and a reduced distribution to fat and muscle. Furthermore, younger age (as in the neonate) will speed the elimination of more soluble anesthetics such as halothane to a greater extent than the less soluble anesthetics, desflurane and sevoflurane. Halothane, and to a far lesser extent isoflurane (1.5 %) and sevoflurane (5 %), undergoes hepatic metabolism. Halothane is reported to undergo as much as 20–25 % metabolism, but at typical anesthetizing concentrations, hepatic halothane removal is extremely small [83].

Renal elimination of drugs and their metabolites occurs primarily by two processes: glomerular filtration and tubular secretion. Glomerular filtration rate (GFR) is only 10 % that of mature value at 25 weeks, 35 % at term and 90 % of the adult GFR at 1 year of age [22, 84]. Aminoglycosides are almost exclusively cleared by renal elimination and maintenance dose is predicted by PMA because it predicts the time course of renal maturation [85]. The kidney is also capable of metabolizing drugs; CYP2E1, which metabolizes ether anesthetics, is active in the kidney. The very presence of CYP2E1 in the kidney has been held responsible for the degradation of ether anesthetics and the release of nephrotoxic fluoride [86].

Immaturity of the clearance pathways can be used to our advantage when managing apnea after anesthesia in the preterm neonate. N₇-Methylation of theophylline to produce caffeine is well developed in the neonate, whereas oxidative demethylation (CYP1A2) responsible for caffeine metabolism is deficient. Theophylline is effective for the management of postoperative apnea in the preterm neonate, in part because it is a prodrug of caffeine, which is effective in controlling apnea in this age group and can only be cleared slowly by the immature kidney [87].

Milrinone, an inodilator, is used increasingly in children after congenital cardiac surgery. Renal clearance is the primary route of elimination. A clearance of 9 L/h/70 kg is reported in adults with congestive heart failure, and we might anticipate that clearance in preterm neonates is reduced to 10 % of this rate because renal function is correspondingly immature in this cohort. This has been confirmed in 26-week PMA preterm infants whose milrinone clearance was 0.96 L/h/70 kg [88]. Similarly the clearance of the neuromuscular-blocking drug (NMBD) d-tubocurare can be directly correlated with GFR [89]. Some drugs such as the nonsteroidal anti-inflammatory drugs (NSAIDs) may compromise renal clearance in early life: ibuprofen reduces GFR by 20 % in preterm neonates, independent of gestational age [90, 91].

The hypnotic actions of propofol result from its interaction with the GABA_A receptor [110, 111].

Integrated PK–PD studies in neonates are lacking, partly due to a lack of consistent effect measures. Consequently, the target concentration for anesthesia in neonates is unknown. The equilibration half-time (T _1/2keo) for the effect compartment is unknown but is assumed to be smaller [109] than the 3 min described in adults [112, 113]. Reduced numbers of GABA_A receptors in the neonatal brain may contribute to a reduced target concentration to effect a response, but this hypothesis remains untested. A circadian night rhythm effect has been noted in an investigation of infant propofol sedation after major craniofacial surgery [114], but such an effect is unlikely in neonates who do not have established day/night sleep cycles.

Propofol is metabolized in the liver with an extraction ratio of approximately 0.9. Clearance is limited by the hepatic blood flow and is thus reduced in children in low cardiac output states [73]. Clearance is affected primarily by UGT1A9 (glucuronidation), with contributions from CYP2B6, CYP2C9, and CYP2A6 isoenzymes, resulting in a more rapid maturation profile than expected from glucuronide conjugation alone (Fig. 3.7).

Although propofol is widely used for target-controlled infusion (TCI) anesthesia in children, commonly used TCI data sets [113, 115–117], have only investigated propofol PK in children beyond infancy. In an effort to link neonatal data with those from children [118], Allegaert used allometry and the Hill equation [62] to suggest a maturation half-time of 44 weeks and a Hill coefficient of 4.9 [119]. Clearance at 28 weeks’ gestation is only 10 % that of the mature value and by full term reaches 38 %. Clearance in the full-term neonate achieves 90 % of the adult value (1.83 L/min/70 kg) by 30 weeks’ PMA. While postmenstrual age (PMA) is the major descriptor of maturation, it is possible that postnatal age (PNA) may also have an additional effect on maturation of propofol clearance above that predicted by PMA [69]. Further longitudinal data that examine individual neonates as they grow are required to clarify this aspect of maturation.

Propofol is used for intubation by neonatologists [120, 121] and anesthesiologists [122, 123]. While doses of 2–3 mg/kg are reported [123–125], caution is advocated in early postnatal life where a transient return to fetal circulation is possible (“flip-flop” phenomenon) due to reduced systemic vascular resistance concomitant with increased pulmonary resistance (associated with hypoxemia and acidosis) [126, 127]. Profound low cardiac output state, together with profound oxygen desaturation that was refractory to usual resuscitation measures including most inotropes, has been reported [125, 128, 129]. Additionally, hypotension of 30-min duration has been reported in preterm neonates given propofol 3 mg/kg for procedural sedation in a neonatal intensive care unit [130], although the severity of the hypotension was similar to that reported after volatile anesthetics at 1 MAC [118], Whether these episodes are attributable to hypovolemia or persistent fetal circulation is not clear, although one study was undertaken in preterm infants within a few hours after birth [125]. In other reports where propofol was administered to older preterm infants, hypotension was not detected [121, 131]. We recommend loading these infants with 10–20 mL/kg balanced salt solution before administration of propofol and atropine 0.02 mg/kg IV and preoxygenation if the risk of bradycardia and desaturation during tracheal intubation are possible. Other adverse effects (bradycardia, propofol infusion syndrome, respiratory depression, immune function) are poorly documented in neonates and require further investigation. Neonates can experience pain on injection and the use of lignocaine to ameliorate this effect is suggested.

Thiopental is an analogue of pentobarbitone. The greater lipid solubility of thiopental is achieved by substituting a sulfur atom in place of an oxygen atom on the barbiturate acid ring [132]. Greater penetration of the BBB has been described in neonates compared with older animals, possibly attributable to the greater blood-brain flow in this cohort [133]. The most likely mechanism of action of thiopental is via binding to GABA_A receptors, which increases the duration of GABA-activated chloride opening.

The ED₅₀ of thiopental varies with age: 3.4 mg/kg in neonates, 6.3 mg/kg in infants, 3.9 mg/kg in children aged 1–4 years, 4.5 mg/kg in children 4–7 years, 4.3 mg/kg in children 7–12 years and 4.1 mg/kg in adolescents aged 12–16 years [134, 135]. It remains to be determined whether altered PK or PD responses explain the reduced dose requirements in neonates. The effect site concentration of thiopental for induction of anesthesia in neonates may be less than that in infants because the neonate has relatively immature cerebral cortical function, rudimentary dendritic arborizations and relatively few synapses [136]. However, integrated PK–PD studies with thiopental in neonates have not been performed to support or refute this premise [137]. The plasma concentration (EC₅₀) of thiopental required for induction of anesthesia in adults based on the EEG is 17.9 mcg/mL; comparable data in neonates are lacking [138]. The T _1/2keo in adults is 0.6 min [138], but there are no estimates in neonates. Children aged 13–68 months given rectal thiopental (44 mg/kg) 45 min prior to surgery were either asleep or adequately sedated with plasma concentrations above 2.8 mcg/mL [139].

Peak concentrations of thiopental are reached in the brain and other well-perfused organs within one circulation time. Recovery is due to redistribution. Reported pharmacokinetic parameter estimates have been derived from infusions administered for seizure control in neonates suffering hypoxic-ischemic insults. Clearance estimates in neonates ranged from 66 to 320 mL/h/kg with a volume of distribution at steady state (Vss) of 3.6–5.4 L/kg [140–143]. Interindividual variability was considerable [137]. While most clearance estimates are less than those in adults (200 mL/h/kg) [138], interpretation is difficult because the hypoxic-ischemic insult will also affect the clearance. Clearance is through oxidation (CYP2C19) to an inactive metabolite, thiopental carboxylic acid, and neonatal immature hepatic function decreases oxidizing capacity. CYP2C19 microsomal activity is approximately 30 % of mature values in the third trimester and increases dramatically around birth [50]. A recent analysis of thiopental clearance maturation is consistent with the timing of maturation of CYP2C19. Clearance rapidly increases during the neonatal period from 33 mL/h/kg at 24 weeks’ PMA to 160 mL/h/kg at term; an adult clearance of 200 mL/h/kg is reported [144]. Neonates (25.7–41.4 weeks’ PMA) undergoing surgery on the first day of life yielded an elimination half-life of 8 h (interquartile range (IQR), 2.5–10.8) and a clearance of 92 mL/min/kg (IQR 20 100) [145]. Thiopental has a low hepatic extraction ratio (0.3), exhibiting capacity-limited elimination. In adults, 10–12 % of thiopental is metabolized per hour; comparable data are not available in neonates. Michaelis-Menten kinetics are reported after prolonged infusion in adults. Michaelis-Menten kinetics are also reported in neonates. The Michaelis constant (Km 28.3 mg/L) is similar to that reported for adults (26.7 mg/L). The maximum rate of metabolism (Vmax) increases from 0.44 mg/min/kg at 24 weeks’ PMA to 5.26 mg/min/kg at term; an adult Vmax of 7 mg/min/kg is reported [144].

These are similar to those described for propofol. Thiopental has little direct effect on vascular smooth muscle tone. Cardiovascular depression is centrally mediated by inhibition of sympathetic nervous activity and direct myocardial depression through effects on trans-sarcolemmal and sarcoplasmic reticulum calcium flux [146]. There is no pain on injection. Because the action of thiopental is terminated by redistribution and metabolism is slow, recovery may be very slow after an infusion of thiopental.

The analgesic properties of ketamine are mediated by multiple mechanisms at central and peripheral sites. The contribution from N-methyl-d-aspartate (NMDA) receptor antagonism and interaction with cholinergic, adrenergic, serotonergic and opioid pathways and local anesthetic effects remain to be fully elucidated.

Ketamine is available as a mixture of two enantiomers; the S(+)-enantiomer has four times the potency of the R(−)-enantiomer. S(+)-ketamine has approximately twice the potency of the racemate. The metabolite norketamine has a potency that is one-third that of its parent. Plasma concentrations associated with hypnosis and amnesia during surgery are 0.8–4 μg/mL; awakening usually occurs at concentrations less than 0.5 μg/mL. Pain thresholds are increased at 0.1 μg/mL [147]. Data from neonates are not available.

Ketamine is very lipid soluble with rapid distribution. Ketamine undergoes N-demethylation to norketamine. Elimination of racemic ketamine is complicated by the R(−)-ketamine enantiomer, which inhibits the elimination of the S(+)-ketamine enantiomer [148]. Clearance in children is similar to adult rates (80 L/h/70 kg, i.e., liver blood flow) within the first 6 months of life, when corrected for size using allometric models [35]. Clearance in the neonate is reduced (26 L/h/70 kg) [149–151], while Vss is increased in neonates (3.46 L/kg at birth, 1.18 L/kg at 4 years, 0.75 L/kg at adulthood [135]). This larger Vss in neonates contributes to the observation that neonates require a fourfold greater dose than 6-year-old children to prevent gross motor movement [152]. There is a high hepatic extraction ratio and the relative bioavailability of oral, nasal, and rectal formulations is 30–50 %.

Ketamine can cause psychotic reactions and hallucinations that can cause distress in older children. These effects can be attenuated by benzodiazepines. An antisialagogue may be effective to diminish copious secretions that may occur after parenteral administration. Tolerance in children may occur with repeated use. Hypocapnia may attenuate ketamine-induced increases in intracranial pressure.

Ketamine is infrequently used in neonates, with perhaps the exception of those with right-to-left congenital heart defects. However, even this limited use has come under scrutiny over concerns that NMDA antagonists (e.g., ketamine and GABA_A agonists) cause significant neuronal apoptosis and other cellular dysgenesis during the period of rapid synaptogenesis in newborn animals [153–155]. Neonatal rats that were anesthetized with ketamine but did not undergo any surgical stimulation or inflammatory response sustained widespread neuronal apoptosis and long-term memory deficits.[156, 157]. The severity of these findings appeared to depend on both the dose and duration of exposure and the coadministration of other proapoptotic compounds in these 7-day-old rats. Moreover, these same findings have been reported in other animals including nonhuman primates and in the presence of anesthetics other than ketamine. More recently, interventions have been forthcoming that appear to prevent or attenuate the neurocognitive sequelae from these anesthetics in newborn animals [158–160]. Nonetheless, extrapolating these animal data to the care of human neonates remains contentious [161, 162].

The washout of inhalational agents follows an exponential decay (the inverse of the washin curve) [168]. The order of the washout of the anesthetics parallels their blood/gas solubilities, that is, the anesthetic that is washed out first, desflurane, is the least soluble in blood [168]. The order of washout in children is expected to be similar to that in adults.

Recovery of motor function in rats parallels the washout of inhalational anesthetics from fastest to slowest: desflurane < sevoflurane < isoflurane < halothane [173]. Notably, the rate of recovery increases in parallel with the duration of anesthesia [173]. In studies in which the recovery from two or more inhalational agents was compared, the end-tidal concentrations of the anesthetics were maintained at approximately 1 MAC until the conclusion of surgery, after which the anesthetics were abruptly discontinued [174–176]. In this paradigm, it is not surprising to find that the rates of recovery paralleled the rates of washout, which in turn paralleled the solubilities of the inhalational agents. In clinical practice however, anesthetic concentrations are gradually tapered as the end of surgery approached. This practice may attenuate the differences in the rates of recovery among inhalational agents reported previously.

In neonates, recovery after inhalational anesthesia is somewhat slower than that predicted solely by the washout curves. In part, this area of investigation has been hampered by a lack of consensus on the essential factors associated with emergence in the neonate [177]. Many anesthesiologists have struggled to explain why neonates recover at a slower rate than expected after a brief anesthetic with only inhalational anesthesia. No clear answers to explain this curiosity have been forthcoming.

In 1969, the MAC of halothane was noted to increase as age decreased [178]. The MAC of halothane reached its maximum value (1.1 %) in the youngest age group, which included 2 neonates and 4 infants 1–6 months of age. Based on this MAC measurement, one study cautioned against the use of halothane in neonates because it caused more hypotension than it did in older infants [179]. We postulated at the time that neonates were more sensitive to the myocardial depressant effects of halothane than children. Subsequently, when we determined the MAC of halothane in neonates as a group distinct from older infants 1–6 months of age, the MAC in the former group, 0.87 SD 0.1 %, was 15–25 % less than that in the latter, 1.2 % [94]. Furthermore, when 1 MAC of inhalational anesthetics was administered, the systolic blood pressure and heart rate responses in neonates were similar to those in older infants [94, 180, 181]. The MAC values in full-term neonates for the inhalational anesthetics in use today are 1.60 SD 0.01 % for isoflurane [182], 9.0 SD % for desflurane [180] and 3.3 SD 0.2 % for sevoflurane [181]. In addition, there is evidence that the primary cause of the increased sensitivity of neonates to cardiodepression by inhalational anesthetics may also be explained by immaturity of the cardiovascular system [183]. Using echocardiography, the cardiodepressant effects of halothane and isoflurane in concentrations up to 1.5 MAC in neonates were noted to be greater than in older infants [184]. The increased susceptibility of neonates to the depressant effects of inhalational anesthetics has been attributed, in part, to a decrease in myocardial contractile elements, a decrease in calcium sensitivity of myocardial fibers [183] and an incomplete sympathetic innervation of the heart and vascular system. Evidence also indicates that both pressor and depressor baroresponses are depressed in the presence of isoflurane in the preterm and full-term neonate [185]. These data suggest that the magnitude of the cardiovascular depression by inhalational anesthetics at 1 MAC in neonates is similar to that reported in older infants and that at concentrations in excess of 1 MAC, it may be greater than in older infants.

In order to attenuate the cardiodepressant effects of inhalational anesthetics in neonates, heart rate should be maintained and preload optimized. Neonatal myocardium evolves during the first month after birth, improving in compliance [186]. Because neonates depend on a rapid heart rate to maintain cardiac output, cardiovascular depression, particularly in the presence of halothane, can be reversed or offset in part by intravenous atropine (0.02 mg/kg) [184, 187]. To optimize preload, balanced salt solution or albumen, in a volume of 5–20 mL/kg, should be administered before anesthesia is induced [93, 184]. Neonates who have had cardiorespiratory dysfunction in the presurgical period or were managed in the neonatal intensive care unit often present to the operating room relatively dehydrated because of aggressive diuretic therapy and/or third-space fluid losses. After rehydration but in the absence of chronotropic agents, systolic arterial pressure decreases ≈ 20–25 % at 1 MAC halothane, desflurane, and sevoflurane compared with awake values in neonates, with either no changes or minimal decreases in heart rate [94, 180, 181]. Similar responses have been reported with 1 MAC of these anesthetics in infants 1–6 months of age.

Curiously, the MAC values for sevoflurane in neonates and infants 1–6 months of age are similar 3.3 and 3.2 %, respectively [181]. It remains unclear why the relationship between the MAC of sevoflurane and age in childhood differs from that for the other inhalational anesthetics. Whether the conformational structure of sevoflurane, a methyl isopropyl ether, or some other physicochemical characteristic is responsible for this unique relationship between MAC of sevoflurane and age is unclear.

The MAC of isoflurane decreases steadily in neonates as gestational age decreases to 24 weeks (Fig. 3.8) [93]. Not only did this study characterize the relationship between the MAC for isoflurane in preterm neonates and age, but it also confirmed the notion that neonates as young as 24 weeks’ gestation respond to noxious stimuli in a predictable manner. Although several explanations have been posited to explain the age-dependent change in MAC in the perinatal period including residual effects of placentally transmitted female hormones, central nervous system substance P and maturation of central nervous system, the cause remains speculative.

Several factors moderate the MAC of inhalational anesthetics in infants and children. The MAC of halothane in children with cerebral palsy and severe mental retardation is approximately 25 % less than that in children without disabilities [188].

Whether these same relationships hold true in neonates has not been established. Acute administration of barbiturates and benzodiazepines decreases MAC [189, 190]; chronic administration of similar medications (such as in the case of children who have been treated with anticonvulsants) does not appear to affect MAC [191]. The effects of specific anticonvulsants such as valproic acid and phenytoin on the MAC of inhalational agents require clarification. Adults who are homozygote for melano-cortin for desflurane require ~20 % more anesthetic (6.2 % vs 5.2 %) than for brunettes [192].

The additivity of MAC fractions of inhalational agents (as well as nitrous oxide) is well established. However, the concept of additivity in children does not hold true for all inhalational agents: the additive contribution for nitrous oxide holds true for the MAC of halothane and isoflurane [193, 194] but not for the MAC of sevoflurane and desflurane [180, 181, 195] children 1–3 years, only 24 %, and that of desflurane only 22 %. Whether the same holds true in neonates is unknown. Additional evidence from the MAC response to tracheal intubation supports this attenuated effect of nitrous oxide on the MAC of sevoflurane in children [196]. This differential effect of nitrous oxide on the MAC values of inhalational agents in children and of its effect on the MAC of sevoflurane and desflurane between children and adults have not been explained.

Nitrous oxide is commonly used to supplement general anesthesia. However, nitrous oxide is infrequently used in neonates, particularly preterm neonates, who require emergency surgery and who are at risk of pulmonary or retinal complications from high oxygen tensions. The avoidance of nitrous oxide in neonates who have bowel obstruction or gas-filled closed spaces is well accepted [197]. However, its avoidance in neonates who are at risk of oxygen toxicity is less clear. It has been recommended that the PaO₂ be limited to a maximum value of 80 mm Hg (that is, an SaO₂ of 93 %) to prevent the effects of oxygen toxicity. This value lies at the shoulder of the steep descending portion of the oxyhemoglobin dissociation curve. If nitrous oxide were used to maintain the arterial oxygen tension below this value, then any minor difficulty with the airway could lead to a rapid hemoglobin oxygen desaturation. This potential problem would be mitigated if nitrogen were used instead of nitrous oxide, since the former is 34 times less soluble in blood than the latter. In these circumstances nitrous oxide should be avoided and replaced with an air/oxygen mixture.

The MAC responses to stimuli other than skin incision, i.e., tracheal intubation, LMA insertion, extubation and return of wakefulness (MAC awake), have also been determined in children, although their applicability to neonates has not been established [198]. The MAC response to tracheal intubation is 10–50 % greater than the MAC for skin incision for halothane [199, 200], enflurane [201], and sevoflurane [196, 202, 203], whereas the MAC for tracheal extubation is approximately 10–25 % less than the MAC for isoflurane [204], desflurane [205 and sevoflurane [206, 207]. The MAC response to tracheal intubation during sevoflurane anesthesia in children is attenuated in the presence of adjuvants such as clonidine [203]. None of these effects have been validated in neonates.

For the past decade, the anesthesia literature has been dominated and preoccupied with the effects of anesthesia on neuroapoptosis in the neonatal brain. Neuroapoptosis is discussed further in Chaps. 4 (Anesthetic techniques for Neonatal Anesthesia) and 15 (Anesthetic Complications in the Neonate) and will not be repeated here.

Although cerebral blood flow (CBF) is autoregulated, there are limits in mean blood pressures beyond which CBF is pressure passive. Evidence suggests that neurological sequelae are increasingly likely when the CBF decreases below 20mL/kg/min. However, there is a dearth of evidence to support this notion in neonates and during general anesthesia [208].

CBF to the cortex is 3–4-fold greater than to white matter, in part, explaining the increasing vulnerability to periventricular leukomalacia in the perinatal period.

All potent inhalational agents depress the central nervous system with dose-dependent decreases in the cerebral vascular resistance and the cerebral metabolic rate for oxygen. The decrease in vascular resistance causes a reciprocal increase in CBF that starts at 0.6 MAC [209]. The extent of the increase in CBF, however, depends on the inhalational agent: halothane > enflurane > isoflurane [209]. The effects of sevoflurane on CBF are similar to those of isoflurane; however, the effects of desflurane differ substantively. Desflurane blunts the cerebral autoregulatory response [210, 211]. The net effect of inhalational agents is an increase in the ratio of the cerebral blood flow to metabolic rate, with desflurane increasing the ratio the greatest.

The effects of inhalational agents on the central nervous system in neonates and children have not been fully elucidated. Preliminary data suggest that cerebral blood flow velocity in children varies directly with the end-tidal carbon dioxide partial pressure during halothane anesthesia [212]. Cerebral blood flow velocity increases as the concentration of halothane increases [213], but does not change significantly in the presence of increasing concentrations of isoflurane.

The EEG activity of desflurane and sevoflurane is similar to that of isoflurane [214, 215]. The electroencephalographic (EEG) activity during halothane anesthesia differs substantially from that during sevoflurane anesthesia [216]. In the case of halothane, the EEG is characterized by slow waves superimposed on fast rhythms (α and β waves), whereas in the case of sevoflurane, the EEG is characterized by mainly sharp slow waves. Furthermore, the shift of power of the EEG from low (1–4 Hz) to medium frequencies (8–30 Hz) is greater for halothane than it is for sevoflurane. The unique EEG tracing in the neonate evolves rapidly in the first few months after birth [217]. During the neonatal period, up to 2 % sevoflurane exerts limited effects on the EEG; however, by 3–5 months of postnatal age, the EEG is responsive to increasing sevoflurane concentrations. The clinical relevance of these EEG differences remains unclear at this time, although the BIS responses, which are derived from the EEG, to sevoflurane differ substantively from those of halothane [105, 218]. Since brain monitors are not used in neonates, this issue is moot.

In children who are anesthetized with sevoflurane, both epileptiform activity in the form of myoclonic movement of the extremities and transient spike and wave complexes on EEG have been reported [219–222].

In two children with histories of epilepsy, diffuse spike and wave complexes were noted on EEG at 5 and 7 % inspired concentrations [220]. In a third child without a history of seizures, a 30-s burst of spike and wave complexes was recorded during sevoflurane anesthesia for spinal surgery but recognized only after the event resolved [222]. None of these three children exhibited clinical evidence of seizure activity. To determine whether these involuntary movements could have as their origin a cortical focus, the EEG was analyzed for evidence of seizure activity in children who had been anesthetized with either halothane or sevoflurane [216]. None of the children displayed either clinical or EEG evidence of seizure activity. The EEG patterns for both halothane and sevoflurane were characteristic even though all of the children had been premedicated with midazolam. The association between myoclonic movement and seizures during sevoflurane anesthesia remains tenuous [223]. A single case of a seizure has been reported in a neonate after sevoflurane anesthesia [224]. However, 3–5 % of children experience at least 1 seizure in childhood, with the greatest incidence occurring in infants <1 year of age [225, 226]. Since epileptiform activity does not portend frank seizures, its clinical relevance during sevoflurane is unknown. Recognizing that idiopathic seizures occur far more frequently than seizures resulting from sevoflurane, the source for such seizures should be sought beyond sevoflurane.

Inhalational agents affect the cardiovascular system either directly (by depressing myocardial contractility or the conduction system or by dilating the peripheral vasculature) or indirectly (by affecting the balance of parasympathetic and sympathetic nervous systems and neurohumoral, renal or reflex responses). The cardiovascular responses to inhalational agents in children are further complicated by maturational changes in the cardiovascular system and its responsiveness to these anesthetics [183, 186]. This is particularly a concern in neonates, although few studies have specifically addressed myocardial contractility in the neonate.

Assessment of cardiovascular variables in infants and children presents a challenge for clinicians. Blood pressure may be measured either invasively (arterial line) or noninvasively. Electrocardiography is routinely used in all age groups to detect arrhythmias. In contrast to blood pressure and electrocardiography, measurement of cardiac output and myocardial contractility is much more difficult to quantify in this age group. Two-dimensional echocardiography and impedance cardiometry have been used to estimate cardiac output and myocardial contractility in infants and children [184, 227–229], although the echocardiographic measurements are subject to variability depending on the preload and afterload. Load-independent derived echocardiographic variables (stress-velocity and stress-shortening indices) have since improved the accuracy of echocardiographic estimates of myocardial function and are used with increasing frequency [230].

In neonates, several factors affect the blood pressure responses to inhalational agents including the particular inhalational anesthetic, the dose, the preload status of the infant, the presence of coexisting diseases and the techniques used to measure the systemic pressure (invasive vs noninvasive) and cardiac function (echocardiography). Most studies demonstrated modest, dose-dependent decreases in blood pressure with all of the inhalational agents. At ~1 MAC, systolic blood pressure decreased ~24 % with halothane, 30 % with sevoflurane and 34 % with desflurane [94, 180, 181]. Systolic pressures may further decrease with concentrations up to 1.5 MAC. Few data exist regarding the hemodynamic responses beyond 1.5 MAC. On balance, all of the inhalational agents (in concentrations up to 1.5 MAC) modestly depress the systemic blood pressure with increasing dose.

The effects of inhalational anesthetics on myocardial contractility and cardiac output in the neonate are incompletely understood. Cardiac output in neonates decreases at 1.0 and 1.5 MAC halothane and isoflurane similarly [184]. Knowing the limited effects of sevoflurane and desflurane on myocardial contractility, one might expect that these anesthetics decrease cardiac output and myocardial contractility at 1 and 1.5 MAC less than halothane [231]. The effects of large concentrations of sevoflurane and desflurane on cardiac function in neonates have been difficult given very large MAC values in this age group. Intravenous atropine and a bolus of balanced salt solution restore, in part, the depressed circulation in neonates [187, 232, 233].

The mechanism by which inhalational agents depress myocardial function remains controversial. Studies in both animal and human myocardial cells suggest that the potent inhalational agents, halothane, isoflurane and sevoflurane, directly depress myocardial contractility by decreasing intracellular Ca²⁺ flux. Inhalational agents decrease the Ca²⁺ flux by their action on the calcium channels themselves, the ion exchange pumps and the sarcoplasmic reticulum [183]. Inhalational agents may also attenuate contractility of ventricular myocytes via voltage-dependent L-type calcium channels (which are responsible for release of large amounts of calcium from the sarcoplasmic reticulum) [183, 186, 234, 235].

That neonates and infants are more sensitive to the depressant actions of inhalational agents than older children is supported by experimental evidence of maturational differences between neonatal and adult rat, rabbit, and feline myocardium [235–238]. Structural differences that may account, in part, for the changes in myocardial sensitivity to inhalational agents with age include a reduction in contractile elements, immature sarcoplasmic reticulum and functional differences in calcium sensitivity of the contractile elements, calcium channels and the sodium–calcium pump in the neonatal myocardium [183, 234–236, 238–241]. The determinants of Ca²⁺ homeostasis in ventricular myocardial cells in the neonate depend on trans-sarcolemma Ca²⁺ flux to a far greater extent than on the sarcoplasmic reticulum [183]. This is based upon a growing body of experimental evidence that includes the finding that the concentration of the Na⁺–Ca²⁺ exchange protein in the neonatal myocardium, a protein that regulates trans-sarcolemma flux of Ca²⁺, exceeds that in adult cells by 2.5-fold and that its concentration decreases with age as the concentration of voltage-dependent L-type calcium channel increases [236]. Furthermore, halothane reversibly inhibits the Na⁺–Ca²⁺ exchange protein in immature myocardial cells [236]. The sarcoplasmic reticulum is poorly developed in neonatal myocardial cells, and this finding weighs heavily against the SR being the major source of Ca²⁺ required for myocardial contractility.

The baroreflex response is also depressed in neonates with both halothane [242] and isoflurane [185], albeit to a greater extent with the former than the latter. In view of the greater incidence of hypotension in neonates and infants than older children, an intact baroreflex could offset, in part, the cardiovascular consequences. However, inhalational agents blunt this response, leaving the infant vulnerable to the cardiovascular depressant actions of inhalational agents. Prophylactic anticholinergics and preload augmentation may attenuate the decrease in cardiac output in the presence of inhaled anesthetics.

Inhalational anesthetics vary in their effect on cardiac rhythm. Halothane may slow the heart rate, in some cases leading to junctional rhythms, bradycardia and asystole. These are dose-dependent responses. Three mechanisms have been proposed to explain the genesis of halothane-associated dysrhythmias: a direct effect on the sinoatrial node, a vagal effect, or an imbalance in the parasympathetic/sympathetic tone. It has also been suggested that the etiology of the bradycardia during halothane anesthesia may be a withdrawal of sympathetic tone. Bradycardia is particularly marked in the neonate, presumably because parasympathetic influences predominate over the sparse sympathetic innervation of the myocardium in this age group. Junctional rhythms are also common during halothane anesthesia. Atrial or ventricular ectopic beats are rare except in the presence of hyper- or hypocapnia. In infants anesthetized with halothane, 10 μg/kg atropine increases heart rate ≥50 % and promotes sinus rhythm [243]. This dose of atropine also increases blood pressure in infants ≥6 months of age and children.

Halothane also sensitizes the myocardium to catecholamines, particularly during hypercapnia. It decreases the threshold for ventricular extrasystoles during epinephrine administration by threefold [244–246]. In contrast, isoflurane, desflurane and sevoflurane maintain or increase heart rate during the early induction period of anesthesia [176, 180, 181, 227, 229, 231, 247–250]. When bradycardia occurs in an anesthetized neonate, the primary cause is always hypoxia, even during anesthesia, before other causes such as a direct drug effect are given serious consideration. The ether anesthetics, isoflurane, desflurane and sevoflurane, do not sensitize the myocardium to catecholamines to the same extent as halothane [244, 245, 251]. The mechanism by which the sinus node controls automaticity is incompletely understood but may include K currents, hyperpolarization-activated current and T and L forms of Ca²⁺ currents [183]. Moreover, developmental changes in these channels likely account, in part, for the differential effects of inhalational agents on heart rate with age [239].

Inhalational agents significantly affect respiration in infants in a dose-dependent fashion via effects on the respiratory center, chest wall muscles, and reflex responses. Halothane depresses respiration by decreasing tidal volume and attenuating the response to carbon dioxide [252–254]. This depression is offset, in part, by an increase in the respiratory rate [253, 254]. These ventilatory responses to halothane are age dependent; minute ventilation in infants decreases to a greater extent than in children [255]. In infants and young children, intercostal muscle activity is inhibited at greater concentrations than the diaphragm [252, 256]. This effect is most pronounced in preterm and full-term neonates and infants and when an endotracheal tube is used in place of an LMA [257]. Isoflurane, enflurane, sevoflurane and desflurane depress the ventilatory drive and tidal volume and attenuate the respiratory responses to carbon dioxide [253, 254, 258–264]. The increase in respiratory frequency that follows respiratory depression may not restore minute ventilation to preanesthetic levels.

Sevoflurane depresses respiration to a similar extent as halothane up to 1.4 MAC, but depresses respiration to a greater extent than halothane at concentrations >1.4 MAC in adults [260]. Sevoflurane does not decrease the tone of the intercostal muscles to the same extent as halothane [256, 260]. The compensatory changes in respiratory rate differ among the anesthetics; respiratory rate increases at ≥1.4 MAC halothane, is unchanged with isoflurane but decreases at ≥1.4 MAC enflurane [253, 254]. When 8 % sevoflurane was compared with 5 % halothane, minute ventilation and tidal volume decreased and respiratory rate increased with both agents to similar extents [265].

Although the physicochemical characteristics of the ether series of anesthetics would predict that anesthesia could be induced smoothly and more rapidly with these agents than with halothane (Table 3.1) [165, 169], this has not proved to be the case. All of the methyl ethyl ether anesthetics irritate the upper airway in children resulting in a high incidence of breath-holding, coughing, salivation, excitement, laryngospasm and hemoglobin oxygen desaturation [175, 247, 295–300].

Consequently, most clinicians avoid inhalational inductions with these anesthetics.

In contrast to the irritant airway effects of the methyl ethyl ether anesthetics, the methyl isopropyl ether anesthetic, sevoflurane, is well tolerated when administered by mask to infants and children at any concentration [174, 176, 183, 301–311]. The incidences of coughing, breath-holding, laryngospasm, and hemoglobin oxygen desaturation during inhalational inductions with sevoflurane whether by slow incremental increases in concentration or a single breath are similar to those that occur during inductions with halothane. The observation that the airway reflex responses are infrequent after a single-breath induction with 8 % sevoflurane or 5 % halothane casts doubt on the adage that high concentrations of inhalational agents trigger airway reflex responses [306]. In fact, the induction is so smooth with sevoflurane even in neonates that dialling 8 % sevoflurane in one step is routine to achieve rapid induction of anesthesia in the neonate without triggering airway reflexes. By increasing the inspired concentration in one step, the period of excitement or agitation during induction is minimized.

Both intravenous and inhalational agents have been used for induction of anesthesia in neonates with congenital heart disease. Sevoflurane compares favorably with halothane for induction of anesthesia in such patients scheduled for cardiac surgery and may be preferred because it maintains cardiovascular stability to a greater extent than halothane, but similar to isoflurane [312–314].

Inhalational agents may be degraded via one of several pathways in the presence of most CO₂ absorbents to form several potentially toxic by-products. Enflurane, isoflurane and desflurane (but not halothane and sevoflurane) react with desiccated soda lime to produce carbon monoxide. Both halothane and sevoflurane may be degraded when incubated with some CO₂ absorbents, yielding compounds that are potentially organ toxic (see Canizarro reaction below) [357]. Two new absorbents may address these clinical risks: molecular sieves [358] and absorbents lacking sodium and potassium accelerants (e.g., Amsorb™) [359–361]. Amsorb™ absorbs CO₂ without releasing carbon monoxide or compound A [359, 360]. Carbon dioxide absorbents differ in their composition and, therefore, in their affinity to interact with inhaled agents. Soda lime contains 95 % calcium hydroxide, either sodium or potassium hydroxide, and the balance as water. Baralyme contains 80 % calcium hydroxide, 20 % barium hydroxide and the balance as water. Amsorb™ contains 70 % calcium hydroxide, 0.7 % calcium chloride, 0.7 % calcium sulfate, 0.7 % polyvinylpyrrolidone and the balance as water. Amsorb™ and other nonreactive absorbents contain neither sodium nor potassium hydroxide, compounds added to improve CO₂ absorption efficiency [360, 362].

Carbon monoxide may be released into the anesthetic breathing circuit when isoflurane or desflurane is incubated with a desiccated CO₂ absorbent. The absorbent within a CO₂ canister may become desiccated if dry fresh gas flows through the canister at a rate sufficient to remove most of the moisture (i.e., >5 L/min continuously through the absorbent canister for 24 h or greater while it is not in service). If the circuit reservoir bag is detached from the canister while fresh gas is flowing, then gas flows through the canister and exits through two sites: a smaller fraction of the fresh gas exits through the inspiratory limb of the canister and a larger fraction of gas flows retrograde through the canister and exits at the site where the reservoir bag should be attached. If the reservoir bag is attached to the canister, then less gas flows retrograde through the canister and the time to desiccate the absorbent is markedly prolonged. Once the absorbent becomes desiccated, carbon monoxide may be released into the inspiratory limb of the breathing circuit if one of the methyl ethyl ether inhalational agents (desflurane, isoflurane or enflurane) is administered [363–365]. The magnitude of the carbon monoxide production follows the order from greatest to least: desflurane ≥ enflurane > isoflurane > > halothane = sevoflurane. Other factors that determine the magnitude of the carbon monoxide concentration produced include the concentration of the inhalational agent, the dryness of the absorbent, the type of absorbent, and the temperature of the absorbent [363]. Currently, carbon monoxide is not detectable by most freestanding anesthetic agent analysers, pulse oximetry or blood/gas analysers (with the exception of co-oximeters), although it is detectable by mass spectrometry. The key to this problem is prevention: turn off the anesthetic machine at the end of the day, disconnect the fresh gas hose to the absorbent canister, leave the reservoir bag connected to the canister, and avoid contact between the desiccated absorbent and the three ether anesthetics, desflurane, enflurane and isoflurane. Others have suggested that high-flow anesthesia should be avoided whenever a circle circuit is used to prevent inadvertent desiccation of absorbent. If the absorbent is desiccated, then some recommend “rehydrating” the absorbent, although this too is fraught with potential problems (including clumping of the absorbent) [366]. If in fact, the absorbent is suspected of being desiccated, then the authors recommend replacing the absorbent before introducing an inhalational agent. Alternatives to conventional absorbents such as the molecular sieve and Amsorb™ may very well obviate all reactions that occur with current absorbents [359, 360]. When these same ether anesthetics are incubated with desiccated Amsorb™, carbon monoxide is not produced [359, 360]. The incidence of carbon monoxide poisoning during anesthesia remains an extremely rare complication when soda lime is used as the CO₂ absorbent. In contrast, the incidence of carbon monoxide poisoning is zero with Amsorb™.

Sevoflurane is both absorbed and degraded via the Cannizzaro reaction in the presence of absorbent resulting in five degradation products [367, 368]. Although the degradation of sevoflurane by the absorbent was initially posited to delay the washin of sevoflurane, recent evidence suggests that the quantity of sevoflurane degraded is of little clinical significance to the speed of washin of sevoflurane [369]. Of the five degradation products between sevoflurane and soda lime, compounds A and B appear in the greatest concentrations in the inspired limb of the breathing circuit. Compound A, fluoromethyl-2,2-difluoro-1-(trifluoromethyl) vinyl ether (also known as PIFE), is nephrotoxic in rats at concentrations ≥100 ppm and has an LC50 of 1100 ppm. Compound B, a methoxyethyl ether compound that is minimally volatile at room temperature, is present in closed circuits at <5 ppm and poses little risk to animals and humans. The remaining three metabolites, compounds C, D, and E, are present in such low concentrations in the breathing circuit that they do not merit further consideration. In a low-flow closed circuit model with an inspired concentration of 2.5 % sevoflurane, the concentration of compound A peaks at 20–40 ppm after several hours of anesthesia [369–373]. Studies are conflicting regarding the risks of using fresh gas flows <2 lpm with sevoflurane and in patients with pre-existing renal dysfunction [374]. In children, compound A concentrations are ≤16 ppm after 5.6 MAC·h sevoflurane in a closed circuit with 2 L/min fresh gas flow [375]. Factors that are known to increase the production of compound A include an increase in the inspired concentration of sevoflurane, Baralyme > soda lime and an increase in the temperature of the absorbent [367, 368].

Studies in rats indicate that under low-flow conditions, compound A is nephrotoxic [376–378].

In contrast, similar studies in humans have yielded inconsistent results [370–372, 379]. The mechanism behind compound A-induced nephrotoxicity is believed to be β-lyase-dependent metabolism to nephrotoxic fluorinated compounds, although this has been the subject of intense debate [283, 380, 381]. More recent evidence suggests that differences in the evidence of nephrotoxicity from compound A in rats and humans relate to differences in metabolic pathways between species [382, 383].

At the present time, sevoflurane is the only inhalational agent for which some federal authorities have recommended minimum fresh gas flow guidelines when it is administered in an anesthetic circuit with soda lime. The minimum fresh gas flow guideline of 2 L/min for >1 h of anesthesia is mandated in only five countries worldwide. The remaining countries where sevoflurane is used have no restriction on the minimum fresh gas flow that can be used in a closed circuit.

Acetaminophen is widely used in the management of pain, but lacks anti-inflammatory effects. Acetaminophen has a central analgesic effect that is mediated through activation of descending serotonergic pathways. Prostaglandin H₂ synthetase (PGHS) is the enzyme responsible for metabolism of arachidonic acid to the unstable prostaglandin H₂. The two major forms of this enzyme are the constitutive PGHS-1 (COX-1) and the inducible PGHS-2 (COX-2). PGHS comprises two sites: a cyclooxygenase (COX) site and a peroxidise (POX) site. The conversion of arachidonic acid to PGG₂, the precursor of the prostaglandins (Fig. 3.15), depends on a tyrosine-385 radical at the COX site. Formation of a ferryl protoporphyrin IX radical cation from the reducing agent Fe³⁺ at the POX site is essential for the conversion of tyrosine 385 to its radical form. Acetaminophen acts as a reducing cosubstrate on the POX site and reduces the availability of the ferryl protoporphyrin IX radical cation (Fig. 3.16). This effect can be reduced by the presence of hydroperoxide-generating lipoxygenase enzymes within the cell (peroxide tone) or by swamping the POX site with substrate such as PGG₂. Peroxide tone and swamping explain acetaminophen’s lack of peripheral analgesic effect, platelet effect and anti-inflammatory effects. Alternatively, acetaminophen effects may be mediated by an active metabolite (p-aminophenol). P-Aminophenol is conjugated with arachidonic acid by fatty acid amide hydrolase to form AM404. AM404 exerts its effect through cannabinoid receptors [384].

Acetaminophen is believed to be an effective antipyretic at serum concentrations of 10–20 mg/L, and these concentrations have been extrapolated to those that provide analgesia. However, the often-cited source for the antipyretic serum concentrations of 10–20 mg/L makes reference to an unpublished source without further elaboration [385]. The calculated ED₅₀ for rectal acetaminophen to avoid any supplemental opioids after day-stay surgery is 35 mg/kg [386]. Time delays of approximately one hour between peak concentration and peak effect have been reported [387, 750]. Pain fluctuations, pain type and placebo effects complicate interpretation of the clinical studies. Population parameter estimates of a maximum effect of oral acetaminophen are 5.17 (the greatest possible pain relief (VAS 0–10) would equate to an Emax of 10) and an EC₅₀ of 9.98 mg/L [388]. The equilibration half-time (T _1/2keo) of the analgesic effect compartment was 53 min with a target effect compartment concentration of 10 mg/L associated with a pain reduction of 2.6/10 [387, 750]. Recent evidence using IV paracetamol generated somewhat different estimates of population parameters with a maximum effect of 4.15 pain units and an EC₅₀ of 2.07 mg/L. However, the equilibration half-time (T _1/2keo) of the analgesic effect compartment was almost double that for oral acetaminophen, 1.58 h [389].

There are few data examining acetaminophen analgesia in neonates. The existing data suggest poor analgesic effect after oral and rectal formulations, after painful procedures [390, 391], after circumcision [392], and during heel prick [393]. This is in contrast to the documented analgesic effect in infants and children [386, 388, 394]. It is unclear why the analgesic effect of acetaminophen in neonates is so poor, but it may be attributable to inadequate serum concentrations, type of pain stimulus or assessment tools for the discrimination of pain. Intravenous paracetamol has been successfully used for neonatal analgesia in the perioperative period [395, 396].

The relative bioavailability of rectal to oral acetaminophen formulations (rectal/oral) is approximately 0.5 in children, but the relative bioavailability is greater in neonates and approaches unity [23]. There are two intravenous paracetamol formulations available and care must be taken with choice of formulation [397]. One is an acetaminophen formulation, whereas the other, propacetamol (N-acetylpara-aminophenoldiethyl aminoacetic ester), is a water-soluble prodrug of acetaminophen that can be administered intravenously over 15 min. It is rapidly hydroxylated into acetaminophen (1 g propacetamol = 0.5 g acetaminophen) [67].

Acetaminophen has a pKa of 9.5, and in the alkaline medium of the duodenum, acetaminophen is non-ionized. Consequently, the absorption half-time of the non-ionized form from the duodenum is rapid in children (T _1/2abs 4.5 min) who were given acetaminophen as an elixir [397]. The absorption half-time in infants under the age of 3 months was delayed, (T _1/2abs 16.6 min) consistent with delayed gastric emptying in young infants [20, 21]. Rectal absorption is slow and erratic with large variability. For example, absorption parameters for the triglyceride base were T _1/2abs 1.34 h (CV 90 %) with a lag time before absorption began of 8 min (31 %). The absorption half-life for rectal formulations was prolonged in infants less than 3 months (1.51 times greater) compared with those in older children [66].

Sulfate metabolism is the dominant route of elimination in neonates, while glucuronide conjugation (UGT1A6) is dominant in adults [22]. Glucuronide/sulfate ratios range from 0.12 in preterm neonates of 28–32 weeks’ PCA, 0.28 in those at 32–36 weeks’ PCA [399] and 0.34 in full-term neonates <2 days old [400]. A total body clearance of 0.74 L/h/70 kg at 28 weeks’ PMA and 4.9 (CV 38 %) L/h/70 kg in full-term neonates after enteral acetaminophen has been reported using an allometric ¾ power model [66]. Clearance increases over the first year of life and reaches 80 % of that in older children by 6 months of postnatal age [23, 62]. Similar clearance estimates are reported in neonates after intravenous formulations of acetaminophen with weight accounting for almost 60 % of the variability in clearance with age (Fig. 3.7) [401–403, 751]. The relative bioavailability of the oral formulation is 0.9. The volume of distribution for acetaminophen is 49–70 L/70 kg. The volume of distribution decreases exponentially with a maturation half-life of 11.5 weeks from 109.7 L/70 kg at 28 weeks postconception to 72.9 L/70 kg by 60 weeks [23] reflective of fetal body composition, and water distribution changes over the first few months of life.

The toxic metabolite of acetaminophen, N-acetyl-p-benzoquinone imine (NAPQI), is formed by the cytochrome P450s CYP2E1, 1A2, and 3A4. This metabolite binds to intracellular hepatic macromolecules to produce cell necrosis and damage. Infants less than 90 days’ PNA have decreased expression of CYP2E1 activity in vitro compared with older infants, children, and adults [53, 404]. CYP3A4 appears during the first week after birth [404], whereas CYP1A2 appears later [54, 404]. Neonates can produce hepatotoxic metabolites (e.g., NAPQI), but the reduced activity of cytochrome P-450 in neonates may explain the rare occurrence of acetaminophen-induced hepatotoxicity in neonates.

Two European centers have published intravenous dosing guidelines for acetaminophen in neonates [397, 405] and the drug is widely used by anesthetists in the UK. Preliminary data [402, 406, 407] from neonates suggest that current dosing regimens are safe. It must be stressed, however, that these are preliminary data only. The combined subject numbers were only 239 neonates. These numbers are too few to exclude the possibility of future hepatotoxicity; caution and continued monitoring of neonates given acetaminophen are vital. The two studies had daily doses that varied by a factor of two. Both authors report satisfactory analgesia, and so it is probably safer to recommend the smaller rather than the larger dosing regimen. Clearance variability and the lack of a suitable marker of reduced clearance in the first few days after birth also support a recommendation of the smaller dose [403]. Several reports have noted accidental 10-fold overdoses of acetaminophen to young infants that fortunately resolved without sequelae [408, 409, 748]. Extreme care must be taken when administering this medication to neonates.

Target analgesic plasma concentrations are believed to be 10–20 ng/mL after major surgery in neonates and infants [71, 435]. The concentration required for sedation during mechanical ventilation may be greater (125 ng/mL) [436]. The large pharmacokinetic and pharmacodynamic variability requires that morphine is often titrated to effect using small incremental doses (0.02 mg/kg) in neonates and infants suffering postoperative pain [437]. The effect compartment equilibration half-time (T _1/2keo) for morphine is ~17 min in adults [438] and is estimated to be 8 min in the full-term neonate [439]. The role of morphine for nursing procedures in the neonatal unit remains controversial. Morphine did not reduce pain responses during endotracheal suctioning, despite plasma concentrations up to 400 ng/mL [64]. Intermittent intravenous acetaminophen reduces the morphine requirements in neonates after major surgery [396].

The principle metabolites of morphine, morphine-3-glucuronide (M3G) and morphine-6-glucuronide (M6G), have pharmacologic activity. Contributions to both the desired effect (analgesia) and the undesired effects (nausea, respiratory depression) of M6G are the subject of clinical controversy [440]. It has been suggested that M3G antagonizes morphine and contributes to the development of tolerance [441].

Opioid analgesics including morphine are widely used to control painful responses in neonates. However, evidence from animals suggests that their use may cause long-term adverse sequelae. In humans, the neonatal use of morphine exerted no long-term effects measured in children 8–9 years of age in terms of intelligence, behavior and visual motor function [442].

Morphine is primarily metabolized by the hepatic enzyme UGT2B7 into M3G and M6G. Sulphation and renal clearance are minor pathways in adults but are more dominant in neonates. Clearance is perfusion limited with a high hepatic extraction ratio. Oral bioavailability is ~35 % due to this first pass effect. The metabolites are cleared by the kidney and, in part, by biliary excretion. Impaired renal function leads to M3G and M6G accumulation.

Morphine clearance matures with PMA reaching adult values at 6–12 months [64]. Clearance increases from 3.2 L/h/70 kg at 24 weeks’ PMA to 9.3 L/h/70 kg at 32 weeks’ PMA and 19 L/h/70 kg at term (Fig. 3.7). The volume of distribution is increased in preterm ventilated neonates (190 L/70 kg) compared with full-term neonates whose lungs were not ventilated (122 L/70 kg) [64]. The volume of distribution in term neonates increased from 83 L/70 kg at birth to a mature value of 136 L/70 kg within 3 months. Metabolite (M3G, M6G) clearance increases with age similar to that described for GFR maturation in infants [63] (Fig. 3.17).

Although an infusion of 5–10 mcg/kg/h will achieve a target concentration of 10 ng/mL in a typical term neonate, clearance is affected by PMA and pathology [65, 443]. Recent evidence indicates that morphine infusions of 4-5 mcg/kg/h provides effective analgesia with minimal risk of overdose in neonates <10 days PNA whereas morphine infusions of 11–23 mcg/kg/h provide reasonable analgesia in older neonates and infants [444]. Morphine pharmacokinetic parameters show large interindividual variability contributing to the range of morphine serum concentrations observed during constant infusions. Protein binding of morphine is small in preterm neonates and has minimal impact on the disposition changes with age. The M6G/morphine ratio also changes with age (Fig. 3.18) because of desynchrony between the maturation of metabolite formation and elimination pathways, but the clinical effect of these ratio changes is unclear.

Although morphine is usually administered intravenously to neonates, other routes have been used. A large variability in the analgesic effect of morphine has been observed after rectal administration; this is a major disadvantage of this route. Although this route is commonly used [445], delayed absorption with multiple doses causing respiratory arrest has been reported [446]. Morphine (25–50 mcg/kg) can also be given via the caudal route to neonates [447, 448]. While systemic absorption is slow, morphine spreads within the CSF to the brainstem where it may cause respiratory depression lasting from 6 h to >18 h [449].

Respiratory depression may occur at concentrations of 20 ng/mL [42] in infants and children, but concentration-response relationships in neonates, particularly preterm neonates prone to physiological apnea, are unknown. Hypotension, bradycardia and flushing reflect morphine’s histamine-releasing characteristics. These are more pronounced with rapid intravenous bolus administration [450, 451]. The incidence of vomiting in postoperative children is related to the morphine dose. Morphine doses in excess of 0.1 mg/kg are associated with a >50 % incidence of vomiting in infants and children [398, 452, 453].

It is difficult to quantify this adverse effect in awake healthy neonates who commonly regurgitate after feeding.

Withdrawal symptoms may be observed in neonates after cessation of a continuous morphine infusion for >2 weeks and after infusion periods <2 weeks if the morphine infusion rate is >40 μg/kg/h. Strategies to prevent withdrawal from morphine include the use of neuraxial analgesia, nurse-controlled sedation management protocols, ketamine or naloxone mixed with morphine infusion and alternate agents (e.g., methadone) with lower potential for tolerance [454, 455].

Fentanyl is a potent μ-receptor agonist with a potency 70–125 times greater than that of morphine. A plasma concentration of 15–30 ng/mL is required to provide total intravenous anesthesia (TIVA) in adults, whereas the EC₅₀, based on EEG evidence is 10 ng/mL [456, 457]. Fentanyl has been shown to effectively prevent preterm neonates from surgical stress responses and to improve postoperative outcome [458]. Single doses of fentanyl (3 mcg/kg) can reduce the physiologic and behavioral measures of pain and stress associated with mechanical ventilation in preterm infants [459]. Fentanyl has similar respiratory depression in infants and adults when plasma concentrations are similar [460].

Fentanyl is metabolized by oxidative N-dealkylation (CYP3A4) into nor-fentanyl and hydroxylized. All metabolites are inactive and a small amount of fentanyl is eliminated via the kidneys unchanged. Fentanyl clearance is 70–80 % of adult values in term neonates and, when standardized to a 70-kg person, reaches adult values (approx. 50 L/h/70 kg) within the first 2 weeks of life [35]. Clearance matures with gestational age: 7 ml/min/kg at 25 weeks’ PMA, 10 ml/min/kg at 30 weeks’ PMA and 12 ml/min/kg at 35 weeks’ PMA [461]. Fentanyl’s volume of distribution at steady state (Vss) is ~5.9 L/kg in term neonates and decreases with age to 4.5 L/kg during infancy, 3.1 L/kg during childhood and 1.6 L/kg in adults [30]. This increased Vss results in a smaller blood concentration after bolus administration in neonates and infants. Administration of fentanyl 3 mcg/kg in infants intraoperatively neither depressed respiration nor caused hypoxemia in a placebo-controlled trial [462, 463].

Fentanyl clearance may be impaired with decreased hepatic blood flow, e.g., from increased intra-abdominal pressure in neonatal omphalocele repair [464]. Infants with cyanotic heart disease had reduced Vss and greater plasma concentrations of fentanyl with infusion therapy [465]. These greater plasma concentrations resulted from a reduced clearance (34 L/h/70 kg) that was attributed to hemodynamic disturbance and consequent reduced hepatic blood flow [466].

Fentanyl is widely distributed with a short duration of effect as a result of redistribution to deep, lipid-rich compartments. Fentanyl redistributes slowly from lipid-rich tissues after discontinuation of therapy, resulting in prolonged periods of sedation and respiratory depression. Although CYP3A4 is subject to single nuclear polymorphisms with slow and rapid metabolizers, redistribution rather than clearance is responsible for offset of the drug effect after single-dose therapy. The context-sensitive half-time (CSHT) after a 1 h infusion of fentanyl is ~20 min, which increases to 270 min after an 8-h infusion in adults [467]. While the CSHT is reduced in children [468], there are no data in neonates. Alternative administration routes (e.g., transdermal and transmucosal) have not been studied in neonates [469]. Epidural fentanyl is widely combined with an amide local anesthetic for provision of postoperative analgesia, although there is limited evidence to support such a combination in young children when a therapeutic concentration of local anesthetic is administered [470]. Spread beyond the level of administration is dose dependent, but limited and respiratory depression is uncommon [471, 472].

Neonatal tolerance to synthetic opioids develops more rapidly (3–5 days) than it does with morphine (2 weeks) and diamorphine (>2 weeks) [473]. Fentanyl also has a propensity for muscular rigidity and laryngospasm with doses as little as 2 mcg/kg IV in neonates [474, 475]. Other drugs metabolized by CYP3A4 (e.g., cyclosporin, erythromycin) may compete for clearance and increase fentanyl plasma concentrations.

Respiratory depression (hours) may outlast fentanyl’s analgesic effect (35–45 min) due to the prolonged CSHT and/or recirculation of fentanyl bound to the stomach’s acid medium (up to 20 % of an IV dose) or delayed release from peripheral compartments.

A target plasma concentration of 2–3 mcg/L is adequate for laryngoscopy and 6–8 mcg/L for laparotomy, and 10–12 mcg/L might be sought to ablate the stress response associated with cardiac surgery [481]. Analgesic concentrations are 0.2–0.4 mcg/L. The T _1/2keo is 1.16 min in adults [76], but the neonatal T _1/2keo has not been reported. Analgesic alternatives should be available for when the short-duration analgesic effect from remifentanil has dissipated. Reports of a rapid development of μ-receptor tolerance with remifentanil use are conflicting. Activity at δ-opioid receptors may contribute [482].

Remifentanil is metabolized by nonspecific esterases in tissues and erythrocytes to carbonic acid [483, 484]; these esterases appear to be mature even in preterm neonates [74]. Carbonic acid is excreted through the kidneys. Metabolism is independent of liver and renal functions. Clearance in patients with butyryl-cholinesterase deficiency is unaffected.

Remifentanil clearance can be described in all age groups by simple application of an allometric model [485]. This standardized clearance of 2,790 mL/min/70 kg is similar to that reported by others in children [77, 486] and adults [76, 483]. The smaller the child, the greater the clearance when expressed as mL/min/kg. Clearances decrease with age, with rates of 90 mL/kg/min in infants <2 years of age, 60 mL/kg/min in children 2–12 years of age and 40 mL/kg/min in adults [75, 485, 486]. The steady state volume of distribution was greatest in infants <2 months of age (452 mL/kg) and decreased to 308 mL/kg in children 2 months to 2 years and to 240 mL/kg in children >2 years [77]. Elimination half-life appears to be constant, ~3 to 6 min independent of the age of the patient and duration of the infusion [77, 476]. Intravenous remifentanil doses of 0.25 μg/kg/min appear to be safe and effective in neonates [479, 480], but PK–PD data in this age group remain limited.

Although covariate effects such as cardiac surgery appear to have a muted effect on PK, as is seen with propofol, cardiopulmonary bypass (CPB) does have an impact. Remifentanil dosage adjustments are required during and after CPB due to marked changes in its volume of distribution [487]. Other PK changes during CPB are consistent with adult data in which a decreased metabolism occurred with a reduced temperature [488] and with reports of greater clearance after CPB (increased metabolism) compared with during CPB [486].

Respiratory depression is concentration dependent [489, 490]. Muscle rigidity remains a concern with bolus doses above 3 mcg/kg used for intubation in neonates [491]. The initial loading dose of remifentanil may cause hypotension and [492] bradycardia prompting some to target the plasma rather than effect site concentration when initiating infusion. This hypotensive response has been quantified in children undergoing cranioplasty surgery. A steady state remifentanil concentration of 14 mcg/L would typically achieve a 30 % decrease in MAP. This concentration is twice that required for laparotomy but is easily achieved with a bolus injection. The T _1/2keo of 0.86 min for this hemodynamic effect [493] is less than remifentanil-induced spectral edge changes described in adults (T _1/2keo = 1.34 min) [76, 494]. Neonatal data are very limited.

Because the inhibitory neurotransmitter, glycine, is included in the formulation of remifentanil, it should not be used for spinal or epidural applications [495].