Pain in preterm infants

Preterm infants have enhanced hormonal and physiological responses to pain. Anand et al showed that preterm infants undergoing patent ductus arteriosus surgery who did not receive fentanyl had a significantly higher hormonal stress response compared with those who received fentanyl.5 More recently, MRI studies identified that exposure to pain and stress in neonatal care alters corticospinal tract, and subcortical white matter development with further evidence of reduced volume of the somatosensory thalamus and changes in the cortex, limbic system and basal ganglia, and the white matter that persist into school age.6

Management of pain and agitation is fundamental to good intensive care, yet neonatal pain is inconsistently assessed and inadequately managed.3 A prospective cohort study of 243 European neonatal intensive care units (NICUs), the European Pain Audit in Neonates (EUROPAIN), reported 34% of admissions, including 82% of those who were mechanically ventilated, received some sedation or analgesia.7 A retrospective cohort study from 348 Pediatrix Medical Group NICUs in the USA found that infants received a sedative, analgesic or paralytic only on 33% of days that they were mechanically ventilated.8

Although optimal management of pain and sedation during mechanical ventilation is as yet undiscovered, selective use of continuous infusion of opioids is recommended for infants who are mechanically ventilated for prolonged periods.3 9 These recommendations stand despite the 2021 update of the Cochrane review that concluded that there is uncertainty whether opioids have any effect on reducing pain in mechanically ventilated infants.2 In addition, there may be short-term and long-term adverse consequences of morphine infusion in preterm infants such as respiratory depression leading to prolonged duration of mechanical ventilation and detrimental long-term neurodevelopmental effects. The Cochrane systematic review2 includes one trial that reported neurodevelopment at 18–24 months10: (RR 2.00, 95% CI 0.39 to 10.29; 78 participants); and one study that reported neurodevelopment outcomes at 5–6 years of age11: (RR 1.60, 95% CI 0.56 to 4.56; 95 participants) and concluded the evidence was uncertain. Puia-Dumitrescu et al12 performed a post hoc analysis of infants who participated in the Preterm Erythropoietin Neuroprotection Trial to investigate the effect of exposure to opioids and benzodiazepines on development at 22–26 months’ corrected age. Of the 936 extremely preterm infants 481 (51%) received opioids; 297 (32%) received both opioids and benzodiazepines; and 20 (2%) received benzodiazepines only. Infants exposed to both groups of drugs had more comorbidities compared with those exposed to only one or none with higher adjusted odds of IVH, NEC, severe bronchopulmonary dysplasia, and retinopathy of prematurity, and longer length of hospital stay. Outcomes at 2 years were negatively associated with the combined use of opioids and benzodiazepines even after adjusting for potential confounders. The lowest scores were noted among the infants who received the drugs for more than 7 days (median (IQR) score of 85 (73–97)) compared with 97 (91–107) in infants with no exposure. However, this was not a randomised allocation and the higher prevalence of multiple comorbidities in the group that received drugs makes it difficult to interpret the results.

Despite these uncertainties, opioids are the most popular choice for management of pain and sedation in mechanically ventilated infants. Of the 2142 mechanically ventilated infants included in the EUROPAIN audit, morphine and fentanyl were given to 923 (43%) and 639 (29%) infants, respectively, while 536 (25%) received midazolam.7 Zimmerman et al similarly reported that in 433 587 ventilated-days, fentanyl (17%), midazolam (14%) and morphine (9%) were the most frequently administered drugs.8 They also reported increase in administration of opioids during mechanical ventilation from 5% of ventilated days in 1997 to 32% in 2012. In the UK, the percentage of very preterm infants who receive intravenous morphine increased from 32% in 2010 to 37% in 2017.13

Dexmedetomidine: an alternative agent for pain and sedation in newborn infants

This high use of opioids despite the uncertainty about benefit and harm is possibly due to a lack of alternative agents that could provide adequate and safe sedation and analgesia in newborn infants. Recently, dexmedetomidine, a highly selective alpha-2-adrenergic agonist (1620:1::α2:α1 specificity) with sedative and analgesic effects, has come into use in newborn infants receiving mechanical ventilation.14 Several recent studies have explored the off-label use of dexmedetomidine in newborn infants for sedation and analgesia in NICU settings. A summary of the studies is given in tables 1–4.

Dexmedetomidine, first approved by the USA Food and Drug Administration in 1999, is similarly authorised by the European Medicine Agency15 for use for sedation in adult intensive care unit patients requiring sedation level no deeper than arousal in response to verbal stimulation and for sedation of non-intubated patients before and/or during diagnostic or surgical procedures requiring sedation. It provides analgesia via inhibition of substance P release from the dorsal horns of the spinal cord while the anxiolysis and sedation are believed to be due to the reduction in sympathetic outflow from the locus coeruleus.16 It stimulates the natural sleep pathways and induces activity similar to non-rapid eye movement sleep in adults and children.17 The effects on respiration and airway also mimic natural sleep and therefore maintain spontaneous breathing and upper airway tone.17

Although there are no data to examine the early exposure to dexmedetomidine and long-term neurocognitive function in children, animal studies suggest that dexmedetomidine may have some neuroprotective effects including cell-protective effects under ischaemic conditions and after traumatic brain injury.18 In rodents, cortical and white matter loss after hypoxic ischaemic injury may be prevented by dexmedetomidine.19 In another rat model, brain tissue and cell loss induced by hypoxia-ischaemia were attenuated by dexmedetomidine post-conditioning, an effect that was inhibited by α2-adrenergic antagonists suggesting that the protective effect was mediated by α2-adrenergic receptor activation.20 However, when combined with hypothermia, Ezzati et al found that dexmedetomidine was associated with adverse cardiovascular events and was neurotoxic following hypoxia-ischaemia.21 In a model comparing the effect of ketamine to that of dexmedetomidine on brain cell degeneration and apoptosis in postnatal day-7 rats, dexmedetomidine produced significant damage in the primary sensory regions which was different to the limbic injury induced by ketamine.22 In 10 mechanically ventilated very low birth weight infants born at 25–30 weeks’ gestational age (GA) who received dexmedetomidine infusion (2–30 days of postnatal age), Cortes-Ledesma et al found that after 6 hours of dexmedetomidine infusion (0.1–0.4 μg/kg/min), brain regional oxygen saturation decreased and fractional oxygen tissue extraction increased without any change in HR, mean BP, oxygen saturation or haemoglobin while amplitude integrated electroencaphalogram (EEG) showed a prolongation of interburst interval.23 As the decrease in brain oxygen saturation could not be related to increased brain activity, it suggested a vasoconstrictor effect on cerebral vessels.23 Currently used sedative and analgesics also adversely impact brain development, for example, exposure to midazolam is associated with macrostructural and microstructural alterations in hippocampal development and lower cognitive scores24 while opioid use is linked to reduced cerebellar volume, poorer cognitive and motor outcomes and behavioural problems in infancy.6 No clinical studies have investigated the impact of dexmedetomidine on neurological outcomes or neurodevelopment in preterm or term-born infants.

Pharmacological studies on dexmedetomidine in newborn infants

The evaluation of pharmacokinetics (PK), pharmacodynamics (PD), efficacy and safety of drugs in vulnerable newborn infants is challenging.24 Dexmedetomidine is metabolised via glucuronidation, hydroxylation (via CYP2A6) and N-methylation and primarily eliminated via the urine.25 As these pathways, particularly glucuronidation, are immature in the first year after birth, dexmedetomidine clearance may be decreased but there is a lack of PK-PD studies in preterm infants.17

The therapeutic window for dexmedetomidine in older children is 0.4–0.8 μg/L (optimum level around 0.6 μg/L)26 with a safety measure of maintaining plasma concentrations below 1 μg/L and is assumed to be the same in neonatal studies.24 The clinical applicability of these serum concentrations is not yet determined and surrogate markers of efficacy such as pain and sedation scores are used to determine clinical utility and titrate dosing,27 The drug is available as dexmedetomidine hydrochloride (100 μg/mL or 4 μg/mL),28 usually delivered in a concentration of 0.5–1 μg/mL as a loading dose (LD) over 10–20 min and maintenance dose (MD) as continuous infusion using rate-controlled syringe pumps. In a phase II/III open-label, safety, efficacy and PK trial including 11 centres, Chrysostomou et al investigated three escalating dose regimens (LD followed by MD) of dexmedetomidine mechanically ventilated 28–44 weeks’ GA infants.29 In 42 infants (18 preterm and 24 term-born), they found that additional sedation (midazolam) was required only in 4 infants (all term-born) while additional analgesia (fentanyl or morphine) was required in 17 infants (3 preterms and 14 term-born). Using the Neonatal Pain and Sedation Scale (N-PASS), they found that at all three dose levels, infants had signs of pain or agitation (N-PASS score >3) at 5% time points only. Preterm infants, as compared with term infants, had lower weight-adjusted plasma clearance (0.3 vs 0.9 L/hour/kg) and increased elimination half-life (7.6 vs 3.2 hours). Van Dijkman et al performed a dose-finding study of dexmedetomidine in mechanically ventilated newborn infants who received dexmedetomidine infusion over 24 hours. Initial dosing (0.3 μg/kg/hour) was determined by extrapolation of previously published PK of dexmedetomidine in older children and was subsequently increased to 0.4 μg/kg/hour in the follow-up phase due to higher clearance of the drug. Additional sedation was required in 83% of the participants as indicated by Comfort-neo scores. In seven term infants with hypoxic ischaemic encephalopathy (HIE) and therapeutic hypothermia (TH), McAdams et al found that dexmedetomidine clearance was either comparable or lower but volume of distribution was larger compared with estimates in normothermic infants.30 They recommended doses up to 0.4 μg/kg/hour and that a LD may be needed to achieve therapeutic concentration without a lag. Dose regimens used in published clinical studies of preterm infants and available neonatal formularies are given in table 5.

Clinical reports of the use of dexmedetomidine in preterm infants

Case reports of dexmedetomidine use to sedate a difficult to ventilate 24 weeks’ GA infant16 and others reporting adverse effects30 31 were followed by larger single-centre studies. Use has been reported in extremely preterm infants in the first weeks after birth,16 27 in preterm infants at later postmenstrual ages (PMA) and term-born infants, after surgical procedures,32 33 and in infants undergoing TH for hypoxic ischaemic encephalopathy (HIE).19 34 35 It is also used for sedation during procedures such as MRI.36 37 In an update of medication prescribing patterns in the US NICUs, Stark et al reported that between 2010 and 2018, dexmedetomidine was ninth in the list of drugs that had the greatest relative increase in use and had become the 90th most frequently prescribed medication38 although its use was not reported, at least until 2017, in neonatal care in the UK.13

Dexmedetomidine is used for sedation and analgesia on its own or in combination with other agents with the aim of reducing cumulative exposure to opioids and benzodiazepines. Comparing 48 infants with a mean GA of 25 weeks who received either fentanyl or dexmedetomidine, O’Mara et al found that the dexmedetomidine group required less adjunctive sedation.28 Comparing 39 infants who received dexmedetomidine, (including 28 who concomitantly received opioid infusion) with 39 matched controls who only received opioids, Sellas et al found that there was no difference in the duration of use or the number of doses of opioids but the total dose of opioids received was lower in the group that received dexmedetomidine.32 With an aim to reduce benzodiazepine exposure in infants who were at least 35 weeks’ PMA and needed sedation for mechanical ventilation, Morton et al implemented a quality improvement project with a sedation guideline recommending the use of dexmedetomidine as initial agent or in addition to midazolam. They demonstrated reduction in use of midazolam with fewer infants receiving midazolam, increased midazolam-free days and lower daily midazolam doses in those who received additional dexmedetomidine without any increase in the mean duration of sedation episodes.39 In 38 infants at a median PMA of 38.8 weeks, of whom 30 were postsurgical, Dersch-Mills et al found that the opioid infusion was reduced within 24 hours of initiation of dexmedetomidine in 12 of the 18 infants who were on concomitant opioids.33

As dexmedetomidine does not cause respiratory depression, it is postulated that its use with no or reduced doses of opioids should enable quicker weaning off mechanical ventilation and support extubation. O’Mara et al found a significantly shorter duration of mechanical ventilation in those who received dexmedetomidine compared with those on fentanyl only (mean±SD: 14.4±7.3 vs 28.4±9.9; p<0.0001). They also reported that clinicians extubated 19 of 28 infants who received dexmedetomidine while receiving infusion of the drug at the time of extubation. Due to the early extubation, fewer infants received dexamethasone (4 vs 11) and there were fewer chest X-rays performed (mean±SD: 28±8.8 vs 49±13.8) in the dexmedetomidine group compared with those who received fentanyl.

The main concerns with dexmedetomidine use are bradycardia and hypotension. In the general adult population, bradycardia is reported in 5%–42% and hypotension in 24%–56% of those who receive dexmedetomidine.40 Chrysostomou et al found, in 42 neonates, that the HR decreased by an average of 12%±9% and systolic BP decreased by 14%±12% at 6–7 hours of starting the dexmedetomidine infusion.29 Although mostly reported to be not of clinical significance, dexmedetomidine can induce reduction in HR and BP. Dersch-Mills et al analysed adverse effects in 38 infants, 36 of whom were also receiving opioid infusion and found that 25 had at least one episode of bradycardia (drop of >20 bpm from baseline lasting for >2 hours) including 7 who had a drop of >40 bpm. Hypotension occurred in 15 infants. Dopamine was used concurrently in 17 infants of which 5 were started on dopamine after the initiation of dexmedetomidine. No infants had dexmedetomidine infusion discontinued due to adverse effects.33 Similarly, Sellas et al found that 5 of 39 infants who received dexmedetomidine experienced episodes of bradycardia as compared with 2 of 39 in the no-dexmedetomidine group although all episodes were self-limiting and did not require any intervention.32 O’Mara et al measured BP and HR hourly and did not see any appreciable effect on either in those on dexmedetomidine compared with the infants on fentanyl28 and Morton et al reported that following their initiative to increase use of dexmedetomidine in mechanically ventilated infants, one infant had dexmedetomidine discontinued due to bradycardia without haemodynamic compromise however discontinuing the drug did not resolve the bradycardia.39 Estkowski et al compared neonates in a paediatric intensive care unit with older infants and found that among the 28 neonates, 4 experienced hypotension and 2 had bradycardia as compared with 30 and 55 of 99 infants who had hypotension and bradycardia, respectively. Three neonates required fluid bolus, one had dopamine infusion started after dexmedetomidine and four had dexmedetomidine discontinued due to adverse events.26

There are no reports of additional adverse events due to dexmedetomidine interacting with other drugs when used in neonatal care, however, polypharmacy is common in neonatal practice and dexmedetomidine can, potentially interact with other frequently used drugs such as phenobarbitone (additional central nervous system depression) and others such as, beta-blockers and diazoxide (worsening of bradycardia and hypotension).41

Conclusions and recommendations

Emerging clinical data suggest that dexmedetomidine may be an alternative to opioids and benzodiazepines for managing pain and sedation in mechanically ventilated newborn infants. It may help reduce the exposure to these agents and as it does not cause respiratory depression, it may facilitate quicker extubation. The potential neuroprotective effect on the immature brain could ease the concerns about the impact of pain and stress and that of opioids and benzodiazepines on the developing brain although other reports suggest that dexmedetomidine could also adversely impact the newborn brain. Despite the creeping acceptance of its use in this population, there are no robust data to demonstrate either the efficacy or the safety of this drug in preterm infants. The studies discussed above are observational and mostly retrospective chart reviews. It is encouraging to see robust PK-PD studies focussing on determining the optimal dose regimens.25 The initial results demonstrating PK differences between term and preterm infants suggest that further population PK-PD studies are required. A Cochrane systematic review was registered in 201642 but there are no completed randomised controlled trials (RCTs) or systematic reviews of use of dexmedetomidine in preterm infants. The increasing number of publications of observational studies and use in clinical practice as seen by the availability of neonatal drug formularies for its use demonstrat evolving practice that dexmedetomidine is being used in newborn infants in neonatal units. The data on safety and efficacy are scant and not of high quality and there are no data on long-term neurodevelopmental impact on preterm or term-born infants. We found one registered pilot RCT21 in term infants undergoing TH for HIE and one investigating the use of bolus dose for supporting less invasive surfactant administration in preterm infants (table 6).

The acceptance of dexmedetomidine in routine clinical practice must be preceded by clinical evidence. While there is a suggestion that dexmedetomidine may be the agent that could provide analgesia and sedation with minimal adverse impacts, the available data are not sufficient to ensure that it is safer than the alternatives used in current practice.43 Although all currently used sedatives and analgesics have adverse impact on the preterm brain, new agents must be investigated thoroughly before being accepted as standard practice. We need adequately powered and well-designed RCTs investigating whether dexmedetomidine alone or in conjunction with opioids/benzodiazepines in preterm infants who require analgesia and sedation during mechanical ventilation reduces the need for opioids/benzodiazepine and improves neurodevelopment at 24 months and later as compared with the use of opioids/benzodiazepines alone.