Objective. To determine the frequency of opiate analgesia administration to infants when life support is discontinued and to determine whether infant characteristics, such as birth weight and diagnosis, or the physician's reasons for discontinuing life support influence either the decision to provide opioid agents or the dosages administered.
Methods. We reviewed all 165 deaths in a 3-year period at a university-based level III intensive care nursery. Of the 121 deaths attributable to withdrawal or withholding of mechanical ventilation and/or extracorporeal membrane oxygenation, we ascertained whether opioid analgesics (morphine sulfate [MS] or fentanyl) were administered either concurrent with or after life-support withdrawal and at what doses. We examined whether these end-of-life practices varied according to birth weight, diagnoses, and the reasons documented by the neonatologist for discontinuing life support.
Results. Opioid analgesia was provided to 84% of infants as their life support was either withheld or withdrawn. Infants with necrotizing enterocolitis and major anomalies or chromosomal disorders were more likely to be given opiates than infants with other diagnoses. Birth weight was not different for infants who received opiates compared with those who were not given opiates. Opioid analgesia was provided to all 18 infants for whom physicians documented the patients' suffering as a reason to discontinue life support. Sixty-four percent of infants who received opiates were given doses in the usual pharmacologic range of 0.1 to 0.2 mg/kg MS. Of the 36 infants given more than 0.2 mg/kg MS, all but 2 were receiving ongoing treatment with opioid agents.
Conclusions. In most cases of withholding or withdrawal of life support in critically ill infants, neonatologists provided opioid analgesia to these infants at the end of life, despite the potential respiratory depression of opioid agents in infants whose respiratory support is discontinued.
Analgesic use for sick neonates has been a controversial topic for many years, as physicians have often disagreed about the infant's capacity to feel pain. Recent evidence has established that neonates, including those born prematurely, are capable of manifesting both behavioral and physiologic responses to nocioceptive stimuli (eg, concurrent with or after surgical procedures, heel sticks, venipunctures, or circumcision).1-10 However, little is known about sick infants' experiences of pain and discomfort during routine handling or noninvasive procedures.11 Although infants' responses to acute pain can be lessened with opioid analgesia,12-14 there is yet no consensus regarding treatment of acute pain for neonates receiving neonatal intensive care.15,16
The recognition of pain in neonates depends on physician and nursing perceptions that certain behavioral or physiologic alterations may be attributable to pain.17 Although neonatal pain scales have been developed,18-23 there is no universally accepted technique for assessing pain in infants24; thus, decisions to treat with opioid agents may depend largely on physicians' beliefs and attitudes about analgesia in sick infants.15,16,25
Whether infants feel pain and discomfort during the dying process is not known. Neither is it known whether adults feel pain associated with hypoxemia and air hunger as they die. Nevertheless, recent studies have found that physicians in adult intensive care units frequently administer opioid agents to dying patients whose life support is discontinued, although they recognize that opiate administration might hasten death as a result of respiratory depression.26,27Similar dilemmas arise when critically ill or dying infants have life support withheld or withdrawn, as neonatologists consider the risk of respiratory depression from opioid analgesia versus untreated pain before death.
Withholding or withdrawing life support has been reported to be the most common mode of death in the intensive care nursery (ICN).28 No studies have yet examined neonatologists' practices of administering analgesia to prevent pain and suffering in critically ill infants as life-sustaining therapy is terminated. The purpose of this study, therefore, was to determine the frequency of opiate administration at the time of life support discontinuation in neonates. Additionally, we examined whether the provision of opioid analgesia to dying infants varied by birth weight, diagnosis, or the reasons cited by the neonatologist for discontinuation of life support.
Of the 165 infants who died in the ICN at the University of California, San Francisco between June 1, 1989, and May 31, 1992, 121 infants died after life support was either withheld or withdrawn. These 121 infants were the subjects of this study. In these cases, either additional treatment necessary for survival was withheld, or life-sustaining treatment was electively withdrawn before the infant's cardiorespiratory arrest.
Diagnostic categories were assigned as follows: (1) extremely low birth weight (ELBW) if birth weight was less than 800 g; (2) intracranial hemorrhage of grades III and IV (by the grading system of Papile et al29) or periventricular leukomalacia; (3) necrotizing enterocolitis (NEC) of at least stage IIb (by the classification of Bell et al30); (4) hypoxic-ischemic encephalopathy; (5) intractable respiratory failure; and (6) major congenital anomalies or chromosomaI abnormalities.
At the University of California, San Francisco ICN, all decisions to limit the treatment of ill newborns are made by the parents, after discussions with the neonatologist and other ICN team members. The medical team, led by the neonatologist, makes treatment decisions such as end-of-life opioid analgesia administration. Parents are usually informed of decisions to provide comfort care to their dying infants. Nursing staff advise physicians when they perceive an infant to be in pain and regularly request that opioid analgesia be administered when life support is being discontinued. The use of benzodiazepines or phenobarbital for sedation was not routine in the ICN during the years of this study.
We defined three categories documented by neonatologists as reasons to limit life support: (1) futility, when death was expected or believed to be imminent despite treatment; (2) severe impairment, if there were concerns about long-term disability; and (3) suffering, if there were concerns that treatment would cause the infant to suffer. We reviewed the neonatologists' daily notes to determine whether one or more of these reasons led to the decision to withdraw or withhold life support.
We determined whether opioid analgesia was administered at the time life support was withheld or withdrawn by reviewing physician orders and nursing documentation. Doses were calculated in milligrams per kilogram for morphine sulfate (MS) and in micrograms per kilogram for fentanyl. Fentanyl doses were converted to morphine-equivalent doses by the formula: 1 μg of fentanyl = 100 μg of morphine.
All data were categorical. Descriptive statistics consisted of frequency distributions of the variables.
This study was approved by the Committee on Human Research at the University of California, San Francisco.
There were 108 infants who died after withdrawal and 13 who died after withholding of life support. Of these 121 infants who died after withdrawal or withholding of treatment, 120 complete records were available for review. Opioid analgesia was administered to 101 (84%) of these infants when life support was withdrawn or withheld.
Opioid analgesia was administered to the majority of infants dying after withdrawal or withholding of life support in each diagnostic category (Table). All 12 infants with NEC were treated with opioid analgesia at the time of life support discontinuation. Infants dying with major congenital anomalies were more likely to have received opioid analgesia at the time of life support discontinuation than were other infants (93% vs 75%). In contrast, ELBW infants were less likely than heavier infants to have received opiate analgesia (66% vs 90%). However, there were 6 ELBW infants whose delivery room resuscitations were withheld or withdrawn because of presumed nonviability. If these nonresuscitated infants are excluded, then ELBW infants were no less likely than heavier infants to have received opioid analgesia (83% vs 90%). Similarly, the mean birth weight of infants treated with opioid analgesia (1993 g) was heavier than infants who received no opiates (1438 g) (P = .05). However, there is no longer a birth weight difference if these 6 nonresuscitated infants are excluded. Overall, in every diagnostic category, opioid analgesia was administered to at least 65% of the infants whose life support was withdrawn or withheld.
Infants whose life support discontinuation resulted from physician concerns about suffering were more likely to have received opioid analgesia at the time of treatment termination than were infants whose life support was ended because of physician assessment of futility. All 18 infants whose therapy was withheld or withdrawn because of concerns about the infant's suffering received opiate analgesia, compared with 48 (81%) of 59 infants for whom physicians documented only concerns about the futility of therapy. Otherwise, the proportion of infants whose treatment was discontinued for quality-of-life concerns, ie, poor neurologic prognosis and/or suffering (87%), was similar to that for infants whose treatment was terminated because of futility (81%).
The median dose of MS (or the equivalent fentanyl dose) was within the usual pharmacologic range of 0.1 to 0.2 mg/kg. Of the 100 infants who received an opiate as life support was discontinued, the majority (64%) received morphine doses in the usual pharmacologic range, whereas 36% received doses greater than 0.2 mg/kg MS. Thirty-four (94%) of the 36 infants treated with higher opiate doses had previously been receiving opioid analgesia and thus may have required higher doses because of tolerance or addiction. Two infants, who were not previously receiving opioid agents, were given MS in excess of 0.2 mg/kg. One infant, a thanatophoric dwarf who was treated with 1 mg/kg MS, died 47 minutes after withdrawal of ventilatory support. Another infant with ventilatory-dependent central apnea and generalized neurologic dysfunction received two MS doses, 90 minutes apart, totaling 0.5 mg/kg. This infant died more than 5 hours after MS administration and extubation.
The median time until death from life support discontinuation was 20 minutes for infants who received more than 0.2 mg/kg MS and 18 minutes for infants who received 0.2 mg/kg MS or less. The median time to death was 18 minutes among the infants who received no MS at the time of life support discontinuation. There were no differences in the doses of MS by birth weight, diagnosis, or the physician's documented reasons for the decision to terminate life support.
In our university-based tertiary care nursery, infants who died when life support was withdrawn or withheld usually received an opioid analgesic as life-sustaining therapy was discontinued. In all cases, death ultimately resulted from the discontinuation of cardiorespiratory support (eg, mechanical ventilation, extracorporeal membrane oxygenation, and vasopressors). Most infants administered opioid analgesia when life support was withdrawn or withheld were given doses within accepted pharmacologic or therapeutic ranges. Thirty-six infants, however, received higher doses than the usual 0.1 to 0.2 mg/kg MS.
In each diagnostic category examined, the majority of infants were administered opioid analgesia as ongoing therapy was withdrawn or as additional treatment necessary for survival was withheld. Infants with major congenital anomalies or chromosomal abnormalities were significantly more likely to receive an opioid analgesic compared with infants with other diagnoses. At our institution, all infants dying with NEC were treated with opioid analgesia as life support was discontinued.
The universal administration of opioid agents to infants dying with NEC in our series suggests that neonatologists at our institution may have considered opioid analgesia a necessary treatment for acute abdominal and/or surgical pain. Whereas analgesia was usually provided to dying infants with other diagnoses, 16% of infants dying with other conditions received no analgesia when life support was discontinued. It may be that neonatologists at our institution provided opiates to dying infants more frequently for known painful conditions than for conditions in which pain was less certain.
We found that the provision of opioid analgesia at the time of withdrawal or withholding of life support was also related to the neonatologist's concerns about the infant's suffering. In all 18 cases in which the neonatologist documented concerns about suffering, opioid analgesia was provided to the infant as treatment was terminated. Our reliance on medical record documentation, however, prevented more detailed information on factors that might have led to the physician's assessment of suffering (eg, observations of infant behaviors, parent's perceptions of suffering, or physician's attitudes about pain and suffering).
Physician attitudes about pain and death may actually be the most important determinants of decisions to provide opiate analgesia as life support is discontinued. Physicians' concepts of the infant's capability of pain perception,15 their concerns about liability, and their prior experiences with dying infants likely influence their thinking about pain in dying infants as well as their decisions to provide opioid analgesia. Although some physicians may be concerned that opiates might hasten an infant's death by respiratory depression, others may administer large enough doses to ensure that there is no suffering, accepting the risks of respiratory depression or significant hypotension. From our medical record review, we were unable to determine the factors that led physicians to administer higher opiate dosages to dying infants.
Infants have been found to be more sensitive than adults to respiratory depression from morphine, an effect that has been found at a morphine dose of 0.05 mg/kg.31 Nevertheless, we found no differences in the time until death after the discontinuation of ventilatory support associated with morphine administration. We also found no difference in the time to death when infants who received high doses of morphine at the end of life were compared with infants receiving standard doses. This finding is not surprising, given that 94% of these infants were previously being treated with opiate analgesia and could have been tolerant to the respiratory depressant effects normally associated with higher doses.
We believe that different practices of administering opiate analgesics to infants at the end of life indicate the recent controversies over pain and its management in neonates. Earlier beliefs held that neonates, especially premature infants, had limited capacities to perceive pain.10,15,17,23,32-34 On this basis, some physicians dispute the premise that pain in infants requires treatment.35,36 Other physicians have higher thresholds for pain treatment in neonates than for older children or adults, believing infants' pain perceptions to be limited; alternately, physicians who believe that infants feel pain may nevertheless overlook its treatment in clinical practice.15,16,25,37-40
In recent years, the growing evidence of physiologic and endocrine stress responses of infants to pain has led to wider acceptance that infants should be treated with appropriate anesthesia and analgesia.10,23,34 Although there are no data regarding pain at the time of death, our study shows that neonatologists at our institution were nevertheless willing to treat potential pain in most infants who died as other life-sustaining treatments were discontinued.
As neonatologists continue to care for extremely ill, marginally viable, and severely damaged infants, decisions to limit either futile or inappropriate therapy will inevitably occur, requiring physicians to decide about the provision of comfort to the dying infant. With no established clinical guidelines, neonatologists may feel uncertain and uncomfortable in the face of unresolved ethical and legal dilemmas accompanying opiate administration to dying infants. This study, the first report of opiate administration to infants whose life support is discontinued, should therefore provide a basis for further research and debate of this complex subject.
We believe that the administration of opiate analgesia to prevent unnecessary pain and suffering at the time of life support discontinuation, even with the risk of respiratory depression and hypotension, may represent rational and humane care for dying neonates. Physicians caring for adults have indicated similar beliefs that providing comfort through pharmacologic means should be the paramount treatment objective when life support is discontinued, even if such treatment risks the hastening of death.27 It has been asserted that the ethical imperative to prevent potential pain should be no less for infants than for adults.10 If this standard is to be upheld, then the provision of appropriate analgesia at the end of life for an infant whose life support is discontinued may be equally as compelling as the provision of appropriate anesthesia for neonatal surgery.
Further investigation of end-of-life care in infants is needed to provide physiologic data to shape guidelines for analgesia when intensive care therapies are discontinued. Future prospective studies of attitudes and treatment decisions by neonatologists are also needed to clarify the ethical principles that should underlie clinical practice. The development of clinical guidelines for opiate administration to dying infants is needed to provide standards of care based on physiologic data, physician and parental attitudes about treatment of pain in infants, and physician beliefs about appropriate ethical practices.
This work (S.N.W.) was supported in part by Perinatology Training Grant HD-07162 from the National Institutes of Health.
We gratefully acknowledge the contributions of Drs George Gregory, Sam Hawgood, Joe Kitterman, and Roderic Phibbs for their helpful comments and review of the manuscript. Additionally, we are grateful for the assistance of Mureen Schlueter for her assistance with the computerized database.
- Received November 10, 1995.
- Accepted February 22, 1996.
Reprint requests to (S.N.W.) 333 E Superior St, Suite 404E, Chicago, IL 60611.
- ICN =
- intensive care nursery •
- ELBW =
- extremely low birth weight •
- NEC =
- necrotizing enterocolitis •
- MS =
- morphine sulfate
- ↵Anand KJ, Hickey PR. Pain and its effects in the human neonate and fetus. N Engl J Med. 1987;317:1321–1329. Review
- Craig KD,
- Whitfield MF,
- Grunau RV,
- et al.
- Craig KD,
- Hadjistavropoulos HD,
- Grunau RV,
- et al.
- ↵Schuster A, Lenard HG. Pain in newborns and prematures: current practice and knowledge. Brain Dev. 1990;12:459–465. Review
- ↵Scanlon JW. Appreciating neonatal pain. Adv Pediatr. 1991;38:317–333. Review
- ↵Boreus LO. Pain in the newborn—pharmacodynamic aspects. Dev Pharmacol Ther. 1990;15:142–148. Review
- Pokela ML
- ↵Schechter NL. The undertreatment of pain in children: an overview. Pediatr Clin North Am. 1989;36:781–794. Review
- Dick MJ
- Pain Management Guideline Panel, Agency for Health Care Policy and Research, US Department of Health and Human Services
- ↵Beyer JE, Wells N. The assessment of pain in children. Pediatr Clin North Am. 1989;36:837–854. Review
- ↵Porter F. Pain in the newborn. Clin Perinatol. 1989;16:549–564. Review
- Wall SN,
- Partridge JC
- ↵Fitzgerald M. Development of pain mechanisms. Br Med Bull. 1991;47:667–675. Review
- ↵Owens ME. Pain in infancy: conceptual and methodological issues. Pain. 1984;20:213–230. Review
- Platt MP,
- Anand KJ,
- Aynsley-Green A
- Williamson PS,
- Williamson ML
- ↵Koren G, Maurice L. Pediatric uses of opioids. Pediatr Clin North Am. 1989;36:1141–1156. Review
- Wellington N,
- Rieder MJ
- Copyright © 1997 American Academy of Pediatrics