Chapter Contents
Introduction 419
Nociception in the infant 419
Effects of nociception 420
Secondary effects of nociception 420
Measurement of pain/nociception in the neonate 420
Pharmacological considerations 421
Clinical analgesia 422
Opioids 422
Tolerance and withdrawal 423
Non-opioids 423
Local anaesthetics 424
Environmental and behavioural interventions 426
Introduction
Pain has been described as ‘an unpleasant sensory and emotional experience which is usually associated with tissue damage or described in terms of such damage’. While neonates have a broadly similar type of response to painful stimuli as adults, the presence or absence of pain as a conscious event can never be proven. Infants and neonates have only a limited ability to communicate compared with adults, and their responses can be inconsistent or absent ( ). Much depends on the nature of self-awareness, consciousness and the development of ‘self’ in fetal life ( ; ; ). Given the impossible task of making judgements on the nature of pain perception in the fetus and neonate, the term ‘nociception’ (the anatomical and physiological system of pain sensation) has been felt to be more appropriate.
Despite increasing interest in provision of analgesia to the neonatal population there is evidence that procedural pain is poorly managed ( ). Moreover, this is not a small problem: neonates, particularly those who are unwell, are subjected to a large number of nociceptive stimuli. A total of 500 invasive procedures were recorded in one neonate during a single hospital stay, and this is not unusual ( ). The American Academy of Pediatrics has produced a policy statement on the prevention and management of neonatal pain which emphasises that painful procedures are a very common part of routine neonatal care ( ).
Nociception in the infant
Nociceptive pathways begin to develop as early as 6 weeks’ gestation when dorsal horn cells in the spinal cord have formed synapses with the developing sensory neurons ( ). These sensory neurons grow peripherally to reach the skin and mucosal surfaces by 20 weeks ( ; ). At full term the density of cutaneous nociceptive nerve endings is at least as great as that of the adult ( ). Organisation of the laminar structure of the cells in the dorsal horn, and their synapses, and the appearance of specific neurotransmitter vesicles begins at 13 weeks and is completed by 30 weeks ( ). By this time, nerve tracts associated with nociception are fully myelinated up to the thalamic level ( ). Synaptic connections of the thalamocortical tracts occur at 24 weeks’ gestation ( ), and myelination of the nociceptive thalamocortical radiations is complete by 37 weeks ( ). Other nociceptive tracts may not be fully myelinated until much later ( ), but lack of myelination does not imply lack of function. Synaptic connections of C-fibres do not appear to mature functionally until the third trimester but noxious stimuli can still be transmitted via A-β fibres ( ). Descending inhibitory tracts, which act via inputs into spinal cord cells to suppress the transmission of noxious stimuli, are also not fully functional at term. The lack of descending inhibition from higher centres will tend to increase afferent nociceptive transmission in the spinal cord.
Few neurophysiological or cytochemical studies have been attempted in the human infant. Positron emission tomography has shown that glucose utilisation, and by inference cerebral metabolism, is maximal in sensory areas of the neonatal brain ( ) and that auditory and visual evoked potentials have developed by 30 weeks’ gestation ( ). These data, along with electroencephalograph data ( ), imply a very complex level of integration and maturity within the cerebral cortex by this time. Somatosensory evoked responses are present from 28 weeks’ gestation, although the latency is long, owing to slow peripheral and central transmission. Studies using functional near-infrared spectroscopy (fNIRS) have demonstrated haemodynamic changes related to clinical noxious stimuli (heel lance ( ) and venesection ( )) in the sensory cortex of newborns as young as 25 weeks’ gestation. These responses are diminished during sleep ( ). have also used event-related evoked potentials to demonstrate cortical responses to heel lance.
While nociceptive connections remain immature in the preterm neonate, the larger receptive fields, the immaturity of the descending inhibitory pathways, and the ability of non-C-fibres to transmit nociceptive inputs into the dorsal horn facilitated by subthreshold C-fibre effects give the impression of an underdamped, poorly discriminative system with a potential for much exaggerated responses. This is borne out in the observations of and on the cutaneous withdrawal reflex and other studies which have shown that newborn reactions to painful stimuli can be diffuse, unlocalised or sometimes completely absent ( ). This failure to respond consistently to a standard noxious stimulus can confound attempts to quantify pain using behavioural measurement.
Effects of nociception
The initial effects of nociception can be categorised under physiological, stress and behavioural responses. All of these responses can compromise the physiological stability of a neonate receiving intensive care.
Haemodynamic responses to noxious stimuli occur as early as 18 weeks’ gestation in the human fetus ( ). Neonates undergoing awake nasotracheal intubation have a rise in mean arterial pressure of 57% with a similar rise in intracranial pressure during the procedure ( ). Age-related differences in cardiovascular responses to noxious stimuli have been reported ( ). In a study on lumbar puncture in preterm infants, the less mature babies (<32 weeks’ gestational age) showed the greatest rise in blood pressure during the handling phase rather than during the actual procedure. This was in contrast to the more mature babies, who displayed maximum response during the procedure itself. Other physiological responses that have been investigated in the context of nociception are R-to-R interval and frequency analysis on electrocardiogram ( ), transcutaneous oxygen tension, ventilatory patterns and sweating.
Hormonal responses to noxious stimuli can also be identified in the human fetus and the response obtunded by opioids ( ). Neonatal stress responses to surgery appear to be greater in magnitude and shorter-lived than in older infants ( ; ) and the subsequent nitrogen loss appears to be greater in the younger age groups. Inadequate suppression of these responses during major surgery affects postoperative recovery ( ), but it remains unclear if complete elimination of the responses is desirable either. Some measure of stress response during surgery can be achieved by real-time analysis of blood sugar provided glucose-containing solutions are avoided ( ).
Preterm infants exhibit hypersensitivity and postinjury hyperalgesia following heel lancing, which can be prevented by the application of local anaesthetic ( ). There is also evidence that early tissue injury can cause hyperinnervation (increased branching of nociceptive nerve endings), which persists and may lead to hyperalgesia and allodynia ( ).
The impression that emerges is that even the very preterm infant has complex interneuronal connections capable of integrated responses to tactile or nociceptive input. They have inconsistent responses to external stimuli, which may reflect the late functional connections of sensory afferents (particularly C-fibres) within the spinal cord. However, the combination of larger receptive fields, recruitment of non-nociceptive afferents and reduced inhibitory controls results in ‘underdamped’ responses (long-lasting, exaggerated and poorly localised) once afferent stimuli have achieved central activation above a threshold level. Inconsistency of response may reflect the profound effects that conscious state ( ) and other external responses ( ) have on behaviour.
Secondary effects of nociception
Neonates who experience repeated noxious stimuli can show both short-term hypersensitivity ( ) and longer term persistence of immature pain responses ( ). Preterm infants have been shown to have increased cortical responses to heel lancing than term born controls ( ). At 18 months they have been reported to respond less than normal infants to everyday painful experiences ( ), and at 4–5 years show increased somatisation (an inappropriate expression of psychosocial distress as physical symptomatology). Awake circumcision without analgesia causes irritability, reduced attentiveness and poor orientation that can last longer than the expected duration of pain ( ), and 3–6 months later circumcised infants have exaggerated responses to painful stimuli compared with a matched group who have not been circumcised ( ).
Comforting strategies that reduce the stress of interventional procedures in preterm infants are associated with improved developmental and clinical outcomes ( ). The results of the multicentre NOPAIN trial evaluated the risks and benefits of providing analgesia to ventilated neonates. A low-dose morphine infusion was found to reduce the incidence of neurological complications (intraventricular haemorrhage and periventricular leukomalacia) in ventilated infants when compared with midazolam or dextrose ( ). The subsequent larger NEOPAIN trial raised concerns about the risk of severe intraventricular haemorrhage, periventricular leukomalacia and death in those who received boluses of morphine ( ). Although later analysis of these data demonstrated that the risk of intraventricular haemorrhage was associated with morphine rather than causal ( ), there have been ongoing concerns about the liberal use of opioids and induction of apoptosis in the developing brain ( ), increase of longer term responses to painful stimuli ( ) and an association with self-destructive behaviour in adolescence ( ).
Measurement of pain/nociception in the neonate
Most of the early studies on infant pain measurement were based on a single noxious stimulus such as heel prick (procedural pain) and were primarily developed for research purposes. This was a useful model because it provided a relatively consistent stimulus from which to identify and grade responses. However, the behavioural tools derived from studies on procedural pain have limited applicability to other situations such as postoperative pain and the discomfort from prolonged immobility in the neonatal intensive care unit. There are few validated tools for prolonged pain or discomfort, they require training to increase reliability and are labour-intensive with a low degree of clinical utility. The Bristol Royal Hospital for Children uses a modified observational pain scale that is simple to use, requires minimal training and is feasible in a clinical setting ( Table 25.1 ). The tools available are either unidimensional, using behavioural responses to pain, or multidimensional, using a combination of behavioural, physiological and contextual indicators. Behavioural indicators are more specific than physiological changes in all age groups. Many of these tools are developed from validated techniques used in older children and infants, such as the Children’s Hospital of Eastern Ontario Scale. Scales such as the Premature Infant Pain Profile (PIPP) have adjustments for gestational age and sleep state, as an acknowledgement that these alter the measurable responses to pain ( ). In the paralysed neonate behavioural tools cannot be used and physiological measures such as cardiovascular responses to handling have to be used.
| AGITATION | FACIAL EXPRESSION | MOVEMENT | VENTILATION |
|---|---|---|---|
| 2 = Major | 2 = Grimace/ nasal flare | 2 = Flexed/tense | 2 = Fighting ventilator |
| 1 = Responds to comforting | 1 = Movement | 1 = Appropriate | 1 = Comfortable |
| 0 = No movement | 0 = No movement | 0 = No movement | 0 = Apnoea |
Table 25.2 shows some of the commonly used pain-scoring systems as described by . The Royal College of Nursing (UK) has published an algorithm for selecting the most appropriate pain scale for specific clinical situations, as well as guidelines for recognition and assessment of acute pain in children (see Weblinks).
| MEASURE | VARIABLES INCLUDED | TYPE OF PAIN | PSYCHOMETRIC TESTING |
|---|---|---|---|
| Premature Infant Pain Profile (PIPP) ( ) |
|
| Reliability, validity, clinical utility well established |
| Neonatal Infant Pain Score (NIPS) ( ) |
| Procedural |
|
| Neonatal Facial Coding System (NFCS) ( ) |
| Procedural | Reliability, validity, clinical utility, high degree of sensitivity to analgesia |
| Neonatal Pain, Agitation and Sedation Sale (N-PASS) |
|
| Reliability, validity, includes sedation end of scale, does not distinguish pain from agitation |
| Cry, Requires oxygen, Increased vital signs, Expression, Sleeplessness (CRIES) ( ) |
| Postoperative |
|
| COMFORT scale ( ) |
|
|
|
There is no ‘gold-standard’ pain scale, as it is not possible to correlate scores with preverbal infants’ pain experience. However have demonstrated a positive correlation of the cortical response measured by fNIRS in term neonates with the PIPP score. The correlation is strongest with the facial (behavioural) rather than the physiological components. However, there are several occasions when the newborns had zero PIPP scores when subjected to heel lances, but had measurable cortical responses. Therefore while pain scoring is the best clinical tool available at the present time, it may underestimate the cortical response, especially in the most preterm ( ) and sick infants.
Pharmacological considerations
Classically, neonates are regarded as highly susceptible to drugs, particularly opioids, in terms of both pharmacodynamics (physiological effects of the drug on the body) and pharmacokinetics (how the body handles the drug). However, the limited studies available have shown that there is large individual variability in this population associated with maturity, previous exposure to drugs and organ function. Individual drug effects from dosing regimens are therefore poorly predictive and can only be described in general terms.
Pharmacodynamics
Opioid receptors change both in numbers and in receptor type during development and in the rat this is associated with a large change in sensitivity ( ). It has been suggested from human studies that have compared plasma concentrations of opioid drugs that human neonates could be relatively resistant to the ventilatory effects of opioids, and that the sensitivity observed after opioid administration is due to selective distribution of the drug to the brain after administration. Lipid-soluble drugs, such as fentanyl, are preferentially redistributed into the neonatal brain after a bolus injection and attain high initial peak concentrations at the effect site (biophase). In contrast, elimination of the drug from biophase is slow because of the limitations on peripheral uptake and drug elimination. Fat-soluble opioids such as fentanyl will therefore have a more rapid onset of effect, greater potency and slower offset than can be predicted by simply analysing pharmacokinetic data. Delivery of morphine into biophase may also be enhanced in the neonate and young infant owing to immaturity of the blood–brain barrier ( ).
Pharmacokinetics
Neonates have a high percentage of body water and less fat than older infants. Consequently, relatively large loading doses of water-soluble drugs such as morphine may need to be given over the first few hours of an infusion to achieve adequate plasma concentrations and effect. Subsequent drug elimination in the ‘drug-naïve’ neonate is delayed due to immaturity of hepatic and renal function. Therefore, once steady state has been achieved infusion rates need to be reduced substantially to prevent accumulation.
A fourfold reduction in the elimination half-life of morphine takes place in the first few years of life, with mean values of 7.2 hours below 1 month, compared with 1.7 hours in adults ( ). This is due primarily to the prolonged clearance of the drug. However there is wide individual variability between subjects, particularly in the neonate. The mean elimination half-life in the newborn was measured at 13.9 hours with a standard deviation of 6.4 hours ( ). Pharmacokinetic data for fentanyl show a similar pattern with even greater variability in the premature infant ( ; ; ). Hepatic clearance of fentanyl may be drastically reduced in neonates undergoing intra-abdominal surgery or those with raised intra-abdominal pressure ( ). This has been attributed to the effects of raised intra-abdominal pressure on liver blood flow ( ). The implications for infants undergoing abdominal surgery are clear: some infants will have a sustained effect from doses of opioid that would normally be expected to have a limited duration of action.
Clinical analgesia
Opioids and paracetamol remain the most commonly used analgesics but there is increasing use of ketamine (an N -methyl- d -aspartate (NMDA) antagonist), clonidine (α 2 agonist), local anaesthetics and non-steroidal anti-inflammatory drugs (NSAIDs).
Opioids
Morphine, fentanyl and codeine are the most commonly used opioids in the UK. All can cause ventilatory depression, hypotension, urinary retention and decreased intestinal motility leading to delayed feeding after abdominal surgery.
Morphine remains the historical gold standard with which other analgesics are compared. Morphine has both slower onset and offset than fentanyl after a single dose, but after long-term infusion this effect is reversed because of the shorter terminal elimination half-life of morphine. In the ventilated neonate, an intravenous loading dose (50–150 µg/kg) is required to achieve effective analgesia, followed by an infusion rate between 5 and 20 µg/kg/h. However as tolerance develops, the infusion rate may need to be further increased. Nurse-controlled analgesia (NCA) is a useful technique for control of infant pain. The baby receives a background morphine infusion (2.5–10 µg/kg/h) topped up at appropriate intervals by ‘nurse-controlled’ doses (varying from about 2.5 to 10 µg/kg) according to formal pain assessment.
Morphine should be used with caution in the spontaneously breathing postoperative neonate. Loading doses of 10–50 µg/kg morphine can be given by slow infusion over 15 minutes in conjunction with sedation scores at 5-minute intervals. Once the desired level of comfort has been achieved, the infusion is discontinued, even if the full dose has not been delivered. Additional doses are then given in the same fashion according to regular documented pain scores, thereby maintaining a therapeutic level on an individual titrated basis. NCA can also be used with a low background morphine infusion (1–5 µg/kg/h). All infants under 6 months receiving opioids need to stay in a high-dependency unit for continual monitoring with direct observation, pulse oximetry and ventilatory monitoring.
Fentanyl provides intense analgesia and relative cardiovascular stability. At high doses (50–150 µg/kg as a single injection), it can control pulmonary hypertension ( ). These doses are well above analgesic doses (0.5–10 µg/kg). observed chest wall rigidity in 8 out of 89 neonates following relatively small doses of 3–5 µg/kg fentanyl, all cases responded quickly to either naloxone or neuromuscular blockade. The high lipid solubility of fentanyl and increased skin permeability of the preterm neonate make transdermal administration a feasible route of administration, which is currently being evaluated. Transtracheal fentanyl provides rapid absorption in rabbits but awaits clinical trials in humans ( ).
Remifentanil is eliminated by the action of plasma esterases and consequently is not dependent on immature hepatic metabolic processes. It is cleared rapidly and predictably, thereby making it an attractive drug for short neonatal surgery when rapid recovery is needed ( ). It has been used to good effect for sedation during mechanical ventilation – doses of up to 0.1 µg/kg/min provided adequate sedation in most infants ( ) – and provides good conditions when used with suxamethonium for tracheal intubation. There is emerging evidence of its efficacy and safety when used to provide intraoperative analgesia in major neonatal surgery ( ; ).
Codeine has, reputedly, a lower incidence of opioid-related side-effects. It has been advocated for use in neonatal practice at a dose of 1 mg/kg orally/intramuscularly/rectally. Single-dose administration appears safe but with repeated doses unwanted side-effects do occur ( . Codeine is metabolised to morphine, but there is considerable variability, and in a small genetic subgroup of ‘poor metabolisers’ (9% in UK, 30% in Hong Kong Chinese) codeine has virtually no analgesic effect ( ). This large interpatient variability necessitates caution when using codeine for long-term use in neonates.
Tolerance and withdrawal
While initial doses of opioid infusions needed for analgesia and sedation are low, the dose requirements increase rapidly. Neonates undergoing extracorporeal membrane oxygenation require five times the initial opioid infusion rate by day 6 to achieve the same level of sedation due to a combination of enhanced elimination ( ) and true tolerance ( ). The use of long-term infusions of morphine for sedation alone is debatable: it is better to reserve analgesic drugs for pain relief or use low-dose infusions of morphine in conjunction with other long-acting sedatives such as chloral hydrate or promethazine.
Opioid antagonists (naloxone 4–10 µg/kg) easily reverse opioid side-effects but must be used with caution with neonates receiving opioid infusions as acute antagonism may trigger a syndrome of withdrawal. Withdrawal is characterised by an excitation of the central nervous system, the gastrointestinal system and the autonomic nervous system ( ). Opioid abstinence syndrome can occur after 48 hours of morphine infusion but is more usually observed after 4–5 days. Management includes the use of a reducing opioid regimen (with morphine or methadone), α 2 antagonists (clonidine 3–5 µg/kg 8–12-hourly) and benzodiazepines for anxiolysis. Weaning schedules can last weeks. Preventive methods include the judicious use of opioids combined with formal comfort scores to optimise the rate of opioid withdrawal.
Non-opioids
Paracetamol (acetaminophen) is primarily metabolised by glucuronidation in older children but by sulphation in neonates ( ). Once the main metabolic pathways are saturated paracetamol is oxidated by the cytochrome P450 system to a reactive intermediary compound which is bound to glutathione, but can react with hepatocyte macromolecules in the absence of glutathione. Neonates may have some protection from the hepatic toxicity effects of paracetamol by having greater glutathione stores and slower oxidative metabolism ( ). Paracetamol is a widely accepted treatment for moderate pain in neonates. Current data suggest that its short-term use in term and preterm neonates is safe and efficacious ( ). It has additive effects when combined with opioids thereby allowing lower doses and subsequently lower incidence of side-effects ( ). Loading doses of paracetamol are similar for premature and term neonates (25 mg/kg orally or 35 mg/kg by triglyceride suppository). Subsequent maintenance regimens must be tailored to the maturity of the infant. found that adequate plasma concentrations can be achieved by an oral dose of 25 mg/kg/day in premature neonates at 30 weeks postconception, 45 mg/kg/day at 34 weeks’ gestation, 60 mg/kg/day at term, and 90 mg/kg/day at 6 months of age. Rectal doses must be increased by a third the oral dose to account for the decreased absorption by this route. These regimens may cause hepatotoxicity in some individuals if used for longer than 2–3 days.
Intravenous paracetamol is available but there is much international discussion as to the safe and most efficacious dose when used in the neonatal population. In the UK it is currently licensed for use in infants aged over 10 days, and less than 10 kg, at a dose of 7.5 mg/kg, not exceeding 30 mg/kg/day; however many paediatric anaesthetists are using higher doses ( ). This dosing does not take into consideration maturational changes in paracetamol metabolism or the evidence that intravenous paracetamol can be given to premature infants and those under 10 days if an appropriately reduced dose is administered. In Australia 15 mg/kg 6-hourly is administered to term infants over 10 days’ age with a reduced dose of 7.5 mg/kg 6-hourly in term infants under 10 days ( ). Further investigation is needed to establish the safety of intravenous paracetamol in premature neonates.
NSAIDs have antipyretic and anti-inflammatory properties with no respiratory depressant or sedative side-effects. As with paracetamol, they have an opioid-sparing effect in older children. Concern over potentially serious side-effects has limited NSAID use for analgesia. Current knowledge of their neonatal effect results almost entirely from their use in the treatment of patent ductus arteriosus (PDA). Side-effects of indometacin include oliguria from decreased renal perfusion; necrotising enterocolitis and gut perforation from decreased splanchnic perfusion ( ); and gastrointestinal bleeding from reduced platelet function. Decreased cerebral blood flow may have a preventive effect on intraventricular haemorrhage in preterm neonates ( ). Ibuprofen has been used as an oral analgesic in term neonates and has a lower incidence of side-effects than indometacin when used for treatment of PDA ( ). It should be used with caution in jaundiced patients as it may displace bilirubin from albumin and, at a dose of 5–10 mg/kg, ibuprofen need only be repeated every 12–24 hours as its half-life is prolonged compared with that in adults ( ).
Over the last decade there has been a resurgence of interest in ketamine and clonidine. Ketamine is a centrally acting NMDA receptor antagonist and potent analgesic. It has been used with proven efficacy and safety for the anaesthesia of neonates for some years. It can be given intravenously (0.5–2 mg/kg), rectally (3–8 mg/kg), intramuscularly (2–5 mg/kg) and via the neuroaxial route ( ). It has the advantage of promoting cardiovascular stability, especially in hypovolaemic patients; it maintains respiratory drive and is a bronchodilator. It increases salivation and respiratory secretions and is usually given with an anticholinergic agent (e.g. atropine 10 µg/kg iv) and is often combined with midazolam (e.g. 25 µg/kg) to provide procedural analgesia and sedation. Intravenous doses of 2 mg and intramuscular doses of 3 mg/kg have been found to induce sedation in 45 seconds and 4 minutes respectively ( ). S(+)-ketamine, one of the two enantiomers of ketamine, has threefold greater analgesic potency than racemic ketamine and may prove to have fewer side-effects. It has been used effectively in caudal anaesthesia ( ) but further clinical trials are needed to evaluate its parenteral use in neonates.
Clonidine is an α 2 agonist that is commonly used via the caudal route (1–2 µg/kg) in combination with local anaesthetic drugs. It provides dose-dependent analgesia following oral and intravenous administration. It has the advantage of having less respiratory depression than opioids, although there are case reports of postoperative apnoea in term and preterm infants ( ; ). Oral doses of up to 4 µg/kg 8-hourly have been used in the treatment of neonatal opioid withdrawal ( ).
Pure sedative drugs have an adjuvant role in the treatment of neonatal pain. In combination with analgesics, they may reduce anxiety and stress by promoting sleep. Of the benzodiazepines, lorazepam (20–100 µg/kg) has a longer duration, resulting in a smoother sedative effect ( ). Benzodiazepines have an additive effect on the respiratory depression of opioids. Chloral hydrate, promethazine and triclofos sodium are other popular sedative drugs.
Local anaesthetics
The use of local and regional anaesthesia/analgesic techniques has a significant, albeit specialised, role in the treatment of neonatal pain ( Table 25.3 ). High-dose opioid techniques may be beneficial to the sick neonate by suppressing nociceptive processing and the neural–humoral responses to pain but are complicated by respiratory depression. Conversely, regional analgesia has been shown to be as effective as morphine infusions without causing sedation or ventilatory depression ( ; ), leading to faster recovery times. There are many recent studies to suggest that in experienced hands regional anaesthesia/analgesia is safe and easy with few side-effects ( ; ; ; ). Local anaesthetics can block nociceptive transmission at various sites – topically and by local infiltration to the skin, regionally by nerve blocks or neuroaxially via the caudal, lumbar or thoracic epidural route. Single-shot administration can provide up to 12 hours of analgesia or indwelling catheters can be used for local anaesthetic infusions or repeat bolus.
