ava i lab le a t www.sc iencedi rec t .com

journa l homepage : www.e l sev i er . com/ loca te /s iny

Seminars in Fetal & Neonatal Medicine (2008) 13, 8e15

Infants of borderline viability: Ethical and clinicalconsiderations

Malcolm Chiswick a,b,*

a University of Manchester, UKb Saint Mary’s Hospital for Women and Children, Manchester, UK

KEYWORDSEthics;Extremely low birthweight;Extreme prematurity;Neonatalintensive care;Neurodisability;Parental involvement;Survival;Value of life;Withdrawal of care

Summary The burden of prolonged intensive care for infants of borderline viability and therelatively high disability rate among survivors pose ethical and clinical problems. Bioethicistshave argued that clinical decisions should be based on the infant’s ‘best interests’, balancingthe burden of intensive care including ‘pain and suffering’ against the likely outcome. How-ever, there are so many uncertainties that the ‘best interest’ argument is more helpful in de-fining problems than driving clinical solutions. The parents’ interests are inextricably linkedwith those of their infant and have considerable weight. Parental complaints about deliveryroom care are rarely based on a conflict of ethical opinion. They are more likely due to mis-understanding, confusion and tension among staff and parents as a result of a failure to havein place or to implement agreed protocols. Information given during pre-delivery counsellingcan easily be misunderstood. The condition of the infant at birth and response to bag and maskventilation have an important role in influencing whether to continue intensive care. Subse-quent care in the neonatal intensive care unit (NICU) should be considered as a ‘trial of life’,with the option of withdrawing ventilatory assistance according to the nature and extent ofneonatal complications.ª 2007 Elsevier Ltd. All rights reserved.

Introduction

Improved survival rates of very preterm infants in the past20 years have resulted in life-saving support being offeredto infants of borderline viability. It was only by ‘testing thewaters’ as part of ordinary clinical practice that the notionof borderline viability could be given meaning in terms ofa gestational age range, which for the purposes of this

* Corresponding author. Highclere, Parkfield Road, Altrincham,Cheshire WA14 2BT, UK. Tel.: þ44 0 161 928 8579; fax: þ44 0 161929 5564.

E-mail address: m.chiswick@manchester.ac.uk

1744-165X/$ - see front matter ª 2007 Elsevier Ltd. All rights reservdoi:10.1016/j.siny.2007.09.007

paper is 21e25 weeks. Our ability to provide for their carein a way that results in a proportion of them who wouldotherwise have died, being discharged home with theirparents has outstripped our capacity to manage wider andinseparable issues that have emerged. The burden ofintensive care and, for many of these infants, a legacy ofimpairments have to be balanced in some way against thebenefits of survival. The problem is magnified in so far ascare offered before birth, at birth and in the neonatalperiod is largely based on evidence derived from researchobservations in more mature fetuses and infants. Thus,there are both ethical and clinical dilemmas.

A nihilistic approach is unhelpful because most pregnantmothers want to take home a healthy baby who will develop

ed.

Infants of borderline viability 9

normally, and it is the role of doctors, midwives and nurses tofacilitate this as far as is reasonably possible. Lack ofevidence for many of the treatments offered to fetuses andnewborns of borderline viability has a positive aspect in thatit serves to highlight the direction that multicentre rando-mised controlled trials (RCTs) need to take in the comingyears.1 Ethical dilemmas exist across all branches of medi-cine e indeed the extremely preterm neonate and the geri-atric patient share much in common in terms of the need toconsider the balance between the burden of treatmentsagainst the potential quality of life, all in the context of fam-ily carers, whether they are parents or sons and daughters.

Outcome of births of borderline viability

Much of the variation in published survival data is influ-enced by the extent to which live births are notified, theattitude of obstetricians to threatened delivery, andpolicies with respect to withholding resuscitation at birthand withdrawing intensive care on the neonatal intensivecare unit (NICU).2,3 There is also likely to be population var-iation due to sociodemographic and genetic factors, as wellas differences in obstetric complications.4

The EPICure study, a prospective observational study ofall births from 20e25 weeks of gestation in the British Islesduring a 10-month period in 1995 has relatively largenumbers of recorded live births at 22e25 weeks gestation,compared with other studies conducted in the 1990s.5,6 Thesurvival and neurodevelopmental outcome information issummarised in Table 1. A further study, ‘EPICure 2’, will re-port on outcomes for babies born at less than 27 weeks ges-tation during 2006. The early results of this study areawaited and will help in our understanding of the patternof any trends in survival rates for extremely preterm infantsover a decade. Comparing different populations over timedoes not provide reliable information about trends.

Table 1 Survival and outcome rates of the EPICure study5,6

Completed weeks of gestation

22 23 24 25

Live births (n) 138 241 382 424Admitted to NMU

(n, %)22 (16) 131 (54) 298 (78) 357 (84)

Survival rates (n, %)Per live births 2 (1) 26 (11) 100 (26) 186 (44)Per admission 2 (9) 26 (20) 100 (34) 186 (52)

Assessed at 6 yearsa

(n, %)2 (100) 22 (88) 73 (74) 144 (79)

Severe disabilityb 1 (50) 5 (23) 21 (29) 26 (18)Moderatec 0 9 (41) 16 (22) 32 (22)Mild/no disabilityd 1 (50) 5 (36) 36 (49) 86 (60)

a 1e3 patients at each gestation died before assessment.b Highly dependent on caregivers, e.g. unable to walk, very

low IQ, profound hearing loss, blind.c Reasonable level of independence, e.g. able to walk, below-

average IQ, correctable hearing loss, impaired vision.d Mild learning difficulty, squint (or no disability).

However, cited data7 from the Trent Region, UK, suggestthat from 1996 to 2003 survival increased for infants bornat 24e26 weeks of gestation, but not for those born at 23weeks of gestation. Recently published national data fromFinland indicate that there was no significant change in sur-vival rates among live born infants of 22e26 weeks of ges-tation, although some neonatal morbidities increased,when a cohort born in 1999e2000 were compared with anearlier cohort born in 1996e1997.8

The interpretation of neurodevelopmental follow-updata is even more problematical, as they are influencedby all the factors that impact on survival statistics. Inaddition, the definitions of outcomes are often not consis-tent between publications and the age of follow-up varies.The subject has been reviewed by Marlow.9 Data from a 6-year follow-up of children in the EPICure study are robustand reflect a national population6 (Table 1).

Ethical considerations: the bestinterest concept

Given the uncertainties about outcomes in individual in-fants, is there an ethical framework that might makedecisions about resuscitation at birth and the provision ofintensive care easier? A central theme put forward bybioethicists and one that has acquired considerable author-ity over the years is the concept of acting in the bestinterest of the infant. This helps us to understand theethical and moral dimensions of providing resuscitation inthe delivery room and intensive support for babies ofborderline viability. It makes us focus on the issues thatinfants could well have an interest in, such as the degree of‘pain and suffering’ involved in their care, whether medicalintervention is futile in the circumstances, the likelihood ofsurvival free of serious disability, and the extent to whichappropriate long-term support would be provided withinthe family should there be a legacy of severe disability.

It is at least debatable whether the newborn, who lacksthe ability to reflect and reason, has any interest in living,or having his or her life prolonged by medical science. Thebest interests of the infant are inevitably based on the per-ceptions of others, be they parents or medical staff. Thelikely outcome is a major factor that informs those percep-tions and therein lies a huge problem. Although we may beguided by data on average survival rates and the risk of dis-ability among survivors, we have no reliable markers of out-come for individual infants at the time when decisionshave to be made shortly before or at birth.

The burden of treatment is reflected in invasive pro-cedures, which are often recurrent and superimposed ona range of deprivations resulting from extremely pretermbirth itself. If the notion of ‘pain and suffering’ is one factor indeciding whether an infant will be offered life-support, thenwe need to consider how we can translate this hugelyimportant concept into something meaningful to infants ofborderline viability. Mostly we rely on the pragmatic, andseemingly sensible, approach of putting ourselves in theposition of the infant and reflecting on what it means. Weare also alert for signals such as restlessness and perhapstransient episodes of tachycardia and tachypnoea. In thiscontext, we might manage or at least control the situation by

10 M. Chiswick

using less invasive procedures where possible and by appro-priate use of analgesics and sedatives. The extent to whichthese approaches really make a difference at the margins ofviability is unknown because we are simply scratching thesurface of a wider issue which is the emergence of minimalconsciousness.10

It is very difficult to apply ethical principles when we arenot sure to what they are being applied. The ‘best interest’argument is helpful in exploring the problems, but of lesspractical help in driving clinical solutions. Its value is betterunderstood in a comparative way when it is argued that theinfant’s interests are more important than the interests ofanyone else. This serves to exclude relatively trivial in-terests that others might have in the infant e for example,for some doctors, the medical challenge of saving a life. Itdoes not exclude, however, the parents’ interests, whichare inextricably linked with those of their infant. Theyshare in the burden of their infant’s treatment, as well asbeing responsible for his or her longer-term welfare, and inthese respects, the parent’s interests are as important asthe perceived interests of their infant.

The Nuffield Council on Bioethics, UK, in a landmarkpublication on ethical issues in critical care decisions infetal and neonatal medicine, made recommendationsabout resuscitation and continuing intensive care of infantsborn at the borderline of viability.7 Their recommenda-tions, which are summarised in Table 2, reflect to a largeextent existing practice in the UK. They are helpful insofaras they consider what should be ‘normal practice’ and thecircumstances under which deviations from this practicewould be acceptable from an ethical perspective. Consider-able emphasis is given to the views of the parents, whilestressing the importance of clinical staff working with theparents in making decisions. Implicit in these recommenda-tions is that the infant’s condition at birth is very influentialin decision-making.

Differences of opinion

Informative counselling probably improves the chances ofparents and doctors agreeing on an appropriate course ofaction for infants of borderline viability. Unresolved

conflict about whether resuscitation and intensive care isappropriate is relatively uncommon, especially when par-ents and staff understand that such decisions can rarely beconfidently made before there is an opportunity to examinethe infant at birth.

Some parents, appropriately counselled before birth,may feel that even with a population survival rate of wellunder 10%, full medical support should be given to theirbaby, even if survival carries a burden of intensive care anda risk of disability. Given the uncertainties around pre-dictions in individual infants and the limitations of the ‘bestinterests’ argument described earlier, it would be pre-sumptuous of doctors to argue that their knowledge of theinfant’s ‘best interests’ was superior to that of the parents.When parents want ‘everything done’ and yet their baby’scondition at birth is extremely poor, based on a carefulclinical examination, the notion of futility will often besupported by a lack of response to bag and mask ventilation.For most parents, this approach will be sufficient to satisfytheir request that everything should be done for their baby.

More problematical for clinical staff is when parentshave made a prior decision that they do not want theirinfant resuscitated and yet assessment at birth indicatesonly moderate depression of vital signs, central pinknessand a peripheral circulation that is judged reasonable. Is itacceptable to countermand the parent’s wishes in theseand similar circumstances?

This issue was addressed in a noteworthy case thatreached the Texas Supreme Court.11 The ethical reasoningbehind the Court’s conclusions resonates with similar situa-tions in other countries where infants of borderline viabilityare offered care. Briefly, an infant was live born at 23weeks of gestation, the parents having been given, beforebirth, a very dismal prognosis for their baby. They indicatedthat no heroic measures were to be taken at delivery andrefused to give prior consent for resuscitation. In the event,a neonatologist did attend the birth and concluded that thefemale who weighed 615 g had a reasonable chance of sur-viving and so she was intubated, ventilated and given on-going care. She subsequently suffered a brain haemorrhageand survived with very severe disabilities. An initial lawsuitfor battery and negligence brought against the hospital

Table 2 Summary of recommendations for the resuscitation at birth of babies born at borderline viability7

Gestation(weeks)a

Standard Exceptions

21 No resuscitation(considered as anexperimentalprocedure)

Only as part of research protocol

22 No resuscitation At parents’ request after prolonged and fully informed discussion of the risks,implications, and the likely outcome

23 Could not be defined .precedence [should be given] to parents’ wishes. If left to clinicians, then the clinicalteam should ‘determine what constitutes appropriate care for that particular baby.’

24 Resuscitation Unless parents and clinicians agree in the light of the baby’s condition that it is notin his or her best interests

25 Resuscitation Unless severe abnormality incompatible with any significant period of survivala Completed weeks of gestation.

Infants of borderline viability 11

succeeded on the basis that the resuscitation was per-formed without consent, and substantial damages wereawarded.

That decision was reversed in a Court of Appeal, andsubsequently came to the Supreme Court where theargument centred on the general rule that a doctor cangive treatment without consent in a life-threatening emer-gency as consent is ‘implied’. The parents’ contention wasthat there was no implied consent because they hadrefused consent some 11 h before the birth. The Court didnot accept this reasoning and instead took the view that‘implied consent’ does not really add anything becausethe doctor carried out the emergency treatment in re-sponse to the emergency, not because there was impliedconsent. That is, even if consent were not implied, therewould still be justification for life-saving treatment (if theattending doctor made the decision in the infant’s best in-terests e my comment). More importantly, the Court tookthe view that decisions about resuscitation made beforebirth were speculative and could only be made when the in-fant was assessed at birth. This highlights the importance ofcounselling parents to the effect that decisions made be-fore birth are temporary and are often influenced by theirbaby’s condition at birth.

When things go wrong in the delivery room leading tocomplaint it is rarely due to a conflict of ethical opinion, oruncertainty about the legal implications of clinical practice.Instead, it is more commonly a reflection of inadequatecounselling, uncertainty, staff confusion and tension sur-rounding the infant’s care. Consider the following:

‘‘They said it was too premature to live and so theywouldn’t monitor my labor. It was like pushing outa corpse. Then they wrapped her up and gave her tomy husband who just sat there with her. After a fewminutes he shouted out ‘she’s moving’. The midwifesaid ‘it’s just a twitch - they often do that.’ Anothermidwife came along and shouted out, ‘My God she’sbreathing - get the paed’. They snatched her up andrushed her away. The next 20 minutes were the worstminutes of my life. Then a lady, the baby doctor,came in and her blouse had blood on it. She said thatmy baby hadn’t responded. I still didn’t know if shewas dead. I only found out when the doctor said thatas it was so premature it could be called a miscarriageif I would prefer that.’’’

This statement illustrates the importance of having inplace and implementing an agreed protocol on the man-agement of infants of borderline viability which shouldinclude appropriate counselling of parents before delivery,the need for a detailed examination of the infant afterbirth and arrangements for care of the dying infant. Theseissues will now be addressed.

Counselling parents

Only a small minority of births at the margins of viabilityoccur so precipitously that there is insufficient time for themother to receive counselling beforehand. Perinatal datafrom surveys that include births at the margin of viabilitysuggest that up to two-thirds of such pregnancies have

complications that would bring the mother to the attentionof the obstetrician days or weeks before delivery.5

The main purpose of pre-delivery counselling is to provideparents with information so that they are in a position tocontribute to decisions about fetal and neonatal manage-ment of their baby. Important issues are the influence ofobstetrical expectations of survival on subsequent manage-ment and the possibility of using antenatal information tohelp in the prediction of outcomes.12 Joint counselling by anexperienced obstetrician and neonatologist reduces the riskof parents receiving conflicting information but does noteradicate it. Given the numerous potential contacts be-tween different staff members and parents, it is not surpris-ing that their perception about treatments and likelyoutcomes may be unclear.

The extent to which counselling in practice achieves theambition of putting parents in a position to make decisionswith the support of clinical staff is variable. A qualitativeresearch study to explore how parents and neonatologistsengage showed how a neonatologist may adopt a neutralapproach after giving information, or a directive approachin which they make a proposal which parents can choose toaccept or not.13 Parents who simply received informationtended to feel isolated if the information did not meet theirexpectations. Parents preferred support and engagement inthe decision-making process. A survey of 149 practisingneonatologists in New England, USA, showed that morethan half saw their main role as providing information,whereas far fewer felt their main role was to help parentsbalance the risks and benefits of treatment options.14 Neo-natal complications were discussed more frequently thanlong-term neurodevelopmental risks. In another study,which included motherecounsellor pairs, there was no cor-relation between the mother’s perception and the counsel-lor’s perception of the process.15 In contrast, motherstended to feel that they had received directive counsellingwith little choice in matters of treatment or resuscitation.

Outcome information presented to parents can easily bemisunderstood and it is important to emphasise the un-certainties involved e especially the difficulty in predictingoutcomes for individual infants. Parents may not readilyunderstand that, even if the intention is to resuscitateand give all possible care, many extremely premature in-fants nonetheless die in the delivery room soon after birthand never reach the neonatal intensive care unit (NICU).Another potential source of misunderstanding is whenparents are counselled in a way that links disability andsurvival rates. For example, at borderline viability, thereis only a small difference between rates of survival withminimal or no disability and rates of survival with majordisability expressed per 1000 live births. The majorimpact of mortality numerically outweighs any consider-ation of disability; it is more helpful to express the risk ofdisability using the numbers discharged from the NICU asthe denominator.

Parents sometimes confuse the difference between via-bility (the ability to sustain life) and vitality (being alive).Inexplicit counselling before birth to the effect that the fetusis too immature to survive and that continuous fetal heartrate monitoring would serve no purpose may be viewed as anindication that the baby will be born dead. This problem iscompounded if only a perfunctory examination of the baby is

12 M. Chiswick

made at birth and signs of life are missed or ignored, asillustrated by the mother’s statement made earlier in thispaper. Uncertainty and tension in the delivery room is proneto occur if such an infant unexpectedly shows increasingvigour. Unless there is evidence of intrauterine death, itshould be emphasised to parents that their baby will be bornalive, the state of the infant at birth is rarely predictable,and that clinical assessment of the infant at birth has animportant role in deciding the type of care that is mostappropriate. Thus, decisions made by parents before birthare not necessarily absolute and binding.

Given the uncertainties about outcomes at 23 weeks ofgestation or less, some parents may not wish their baby toreceive medical support at birth, regardless of his or hercondition. It is important in these circumstances for parentsto understand that their baby may nonetheless survive forsome hours or even days and that they need to contributeto the decision about the subsequent care of their baby, asmuch as they contributed to the decision to withholdmedical support.

In view of all the variables implicated in counselling,parents may find it helpful to be presented at the outsetwith a summary of the potential hurdles according tochronological stages. First, the notion of fetal survival tobe live born provides an opportunity to discuss obstetricmanagement. Second, the challenge of survival in the deliv-ery room raises issues around the unpredictability of the in-fant’s condition at birth, the need for assessment, andwhat resuscitation might achieve. Third, care in the NICUraises issues around the burden of intensive care, the con-trol of pain and discomfort, common neonatal complica-tions, and possible reasons for withdrawing assistedventilation. Fourth, issues around the risks and potentialnature of impairments and disabilities can be discussed inthe context of outcome data from published studies, andthe unpredictability of neurodevelopmental outcomes inthe early weeks and months.

Assessment and delivery room care

Infants of borderline viability should be carefully assessedat birth by an experienced neonatologist. The purpose is toconfirm or otherwise signs of life; to form an opinion onwhether the apparent gestational age based on physicalappearance and size is consistent with information beforebirth; to assess the condition of the infant in terms of vitalsigns and the presence of malformations; and to decidewhether any medical intervention is appropriate, takinginto account any decisions that may have been made withthe parents before birth.

Although the definition of ‘viability’ is open to discus-sion, signs of life are either present or not. The WorldHealth Organisation’s definition of live birth is as follows:‘‘Live birth refers to the complete expulsion or extractionfrom its mother of a product of conception, irrespective ofthe duration of the pregnancy, which, after such separa-tion, breathes or shows any other evidence of life e e.g.beating of the heart, pulsation of the umbilical cord ordefinite movement of voluntary muscles e whether or notthe umbilical cord has been cut or the placenta is attached.Each product of such a birth is considered live born.’’

In spite of the potential problems that arise in somecircumstances when this definition is strictly adhered to,there are more important issues at stake when a culture ofignoring signs of life is prevalent in the delivery room(Table 3).

In circumstances of a live birth at 23 weeks of gestationor less, where there is no reason to doubt the gestationalage, and where parents have made an informed priordecision that resuscitation is not to be carried out regard-less of their baby’s condition, then in the light of currentknowledge there is probably no obligation for doctors tocountermand the parents wishes. In these circumstances,the nature of comfort care should accord with the parent’swishes. Some will want to hold their baby for extended pe-riods e others will prefer their baby to be nursed in a cot. Itis inappropriate for such infants to be admitted to the NICUto have their death medically supervised. The use of drugsis mentioned later in the context of withdrawal of ventila-tory support in infants nursed on the NICU.

For most newborns, especially those born at 24e25 weeksof gestation, the condition of the infant at birth andthereafter will influence clinical management. There is noevidence-based guidance on the safe and effective deliveryroom management of infants of borderline viability. There-fore, all management protocols are somewhat empirical.The infant’s condition as assessed by the Apgar score is nota predictor of short- or long-term outcome. However, if wewere to argue that there is no clinical basis for determiningthe appropriateness of resuscitation and continuing medicalsupport, then this would be tantamount to abandoning suchinfants without examining them. Unless the infant hasa serious congenital malformation, or unless there was fetalbradycardia before birth for more than a few minutes,assessing the infant’s response to resuscitation is morallyand probably medically a better alternative. Note, inciden-tally, that even if continuous electronic fetal monitoringwas not considered appropriate, it is helpful for neonatol-ogists to know what the fetal heart rate was shortly beforebirth using intermittent auscultation or a handheld Dopplerdevice.

Endotracheal intubation of extremely small infants canbe difficult even in experienced hands. The prime intentionis to provide effective positive pressure ventilation and thusstabilise the infant. However, the procedure, especiallywhen difficult, can convert a moderately depressed infantinto one who is moribund. In view of the need to minimisediscomfort and harm, a reasonable approach is to assess theinfant’s response to effectively applied bag and maskventilation. In this context, a poor response would beindicated by persisting bradycardia, clinical signs of

Table 3 Implications of erroneously failing to classifybabies of borderline viability as live births

� Generates an inappropriate culture in the delivery room.� Leads to a casual approach in assessment of the newborn.� Distorts birth statistics and outcome data.� Shields the public, health service managers and politicians

from one of the challenges in perinatal care.� May culminate in allegations of negligence.

Infants of borderline viability 13

circulatory impairment and no spontaneous breaths, in thepresence of hypothermia and/or an arterial pH level <7.1.Of course, judgements must be made when some but not allof these features are present. Although this approach islargely empirical, there is some evidence to support it fromthe EPICure study. In a multivariate analysis, this showedthat a heart rate less than 100 beats per minute at 5 min anda body temperature less than 35 �C on admission to the NICUwere indicators of an increased risk of mortality beforedischarge.5 Those infants who do not show a significantresponse to non-invasive resuscitation may be given appro-priate comfort care in accordance with the parents’ wishes.

Limits of continuing care on the NICU

For those who do improve with resuscitation and who aretransferred to the NICU, there are essentially two alterna-tive approaches, neither supported by a very strongevidence-base. Elective intubation has the advantage ofproviding the route for surfactant administration. Respira-tory support may be continued with appropriate positivepressure ventilation or with endotracheal continuous pos-itive airways pressure (CPAP) treatment, depending on theinfant’s respiratory drive.

The alternative for infants with reasonable respiratorydrive from the outset is to support them by nasal CPAP withintubation being used transiently to give surfactant. Aproportion of infants managed in this way will still eventu-ally require mechanical ventilation for prolonged periods.Greisen16 has described the Danish experience of a mini-mally invasive approach to resuscitation and on-goingcare. Deferring mechanical ventilation in this way carriesat least a theoretical risk of harmful effects of recurrentapnoea and respiratory acidaemia. A follow-up study of in-fants born at less than 28 weeks of gestation did not suggestan increased risk of intellectual impairment in infantstreated in this way with CPAP, but relatively few infantsof less than 25 weeks of gestation survived.17

Infants of borderline viability are at great risk ofdeveloping a wide range of neonatal complications, thedetails of which are beyond the scope of this article. Thereis the temptation for some clinical staff to see eachcomplication as a challenge that must be met at all costssimply because the infant is a patient on an NICU. Incontrast, the NICU should be seen as providing a supportiverole to parents while their baby is essentially undergoinga ‘trial of life’, which commenced with resuscitation in thedelivery room. A key concept during this time is the need tominimise as far as possible the burden of intensive carewhile assessing the infant’s progress with the parents atfrequent intervals, with a view to timely intervention towithdraw ventilatory assistance should that be indicated.

Although on ethical grounds there is no differencebetween withholding ventilatory support and withdrawingit, so long as it is done in the infant’s best interests, there isa huge difference from a clinical perspective. The decisionto withdraw assisted ventilation is made only after theinfant has been given a chance of life when there has beentime to serially assess clinical progress and response totreatments. There are nonetheless great difficulties in mak-ing decisions about the withdrawal of life-support, and

many find that it is relatively easier to make a decisionnot to commence it.

The futility of continuing treatments

A treatment such as mechanical ventilation is futile when itserves no useful purpose and may instead be prolonging thedying process rather than offering a reasonable hope ofsaving life. A point of futility is ordinarily reached when it isfelt that a patient has ‘entered the process of dying’. This isnot always easily definable, but it may be portrayed as aninfant with multiple organ failure who, in spite of medicalattempts to treat or control the situation, shows no signs ofimprovement or who steadily deteriorates. Few woulddisagree with the notion of futility in this context.

The challenging question is whether, in the circum-stances that I have described as a ‘trial of life’, there isa lesser standard of futility that is more appropriate. Is it inthe best interests of the infant and is it fair on the parentsthat the burden of intensive care, however we try tominimise it, should continue until the point of dying isreached? Against the background of an uncertain outcomeby virtue of borderline viability, any additional neonatalcomplications that arise during the course of the infant’scare add to both the burden of suffering and the burden oftreatments while doing nothing to improve the prognosis.Thus, withdrawal of care might be considered when aninfant suffers one or more complications that are burden-some, such as moderately severe respiratory distresssyndrome, sepsis, necrotising enterocolitis, seizures orcirculatory impairment not responsive to inotropes. Ineffect, this amounts to a ‘quality of life’ decision appliedto the here and now.

A decision to withdraw ventilatory assistance based onthe predicted future quality of life is barely applicable toinfants of borderline viability because they already carrya significant but unpredictable risk of disability by virtueof their birth. In current clinical practice, the event thatmight convert an unpredictable but significant risk intoone of near-certainty is a massive lesion observed on anultrasound brain scan, such as a very severe intraventricu-lareperiventricular haemorrhage. However, most affectedinfants would have associated clinical signs and symptoms,which by themselves would lead to a consideration of with-drawing life-support.

When the parents agree that ventilatory support shouldbe withdrawn, they should be informed what the processnormally entails, and given the opportunity to contribute totheir baby’s care after extubation. This helps them tounderstand that following extubation a variable length oftime will elapse before their baby will die.

Care after withdrawing ventilatory support

Parents may feel uncomfortable asking how long it will bebefore their baby dies and yet it may be the most pressingquestion on their minds. If the infant had truly entered theprocess of dying before extubation, then it is likely thatdeath will occur very shortly after extubation e perhapswithin 30 min or so. When assisted ventilation is withdrawnfrom infants who cannot be said to have reached the stageof dying then they are more likely to show signs of life for

14 M. Chiswick

an extended period often measured in hours, and some-times even days. For many parents this can be distressing.For others, it is an opportunity to relate to their baby out-side the paraphernalia of intensive care.

Some infants gasp intermittently for prolonged periods(agonal breathing). We cannot know what this means forinfants, but instinctively we feel that it must be mostunpleasant and that it amounts to unnecessary suffering. Itis certainly very stressful for parents and indeed staff.Effective support of parents through this difficult time isessential. This emphasises the importance for NICU staff todiscuss, develop and implement protocols for the care ofdying infants. Infants are often prescribed sedation afterextubation; this is acceptable provided the intention, dose,and the manner of administration are consistent with theprime aim of relieving the infant’s perceived discomfort.

In conclusion, the care of babies of borderline viabilityshould be seen as the provision of clinical care within anethical framework. The notion of acting in the infant’s bestinterests helps clinicians to consider the burden of in-tensive care and futility weighed against likely outcomes.However, many of these issues are uncertain and in thatrespect the best interest concept rarely provides clinicalsolutions. The evidence-base for many of the treatmentsoffered to infants of borderline viability has yet to bedeveloped. It is because of the uncertainties that mater-nity units need to develop and implement agreed protocolsso that babies of borderline viability and their parents arenot met by confusion and crises in the delivery room. Wehave to acknowledge that much of the guidance will beempirical; we may not always get it right, and with hind-sight, we may wish that we had adopted an alternativeapproach.

Practice points

� Survival at borderline viability is achieved aftera prolonged burden of intensive care and, formany infants, there is a legacy of disability. Fac-tors used to assess an infant’s ‘best interests’ areoften shrouded in uncertainty and are of limitedhelp in driving clinical solutions.� The parents’ interests are inextricably linked with

those of their infant and have considerableweight.� When things go wrong in the delivery room leading

to parental complaints, it is rarely due to a conflictof ethics, and is more likely due to staff confusionand tension due to a lack of agreed protocols onbirths of borderline viability.� Parents given a dismal prognosis for their baby be-

fore birth may not expect their baby to be bornalive, unless this is made explicit duringcounselling.� The infant’s condition at birth and response to bag

and mask ventilation are important factors in de-termining whether continuing intensive care isappropriate.

References

1. Higgins RD, Delivoria-Papadopoulos M, Raju TNK. Executivesummary of the workshop on the border of viability. Pediatrics2005;115:1392e6.

2. Hansen BM, Greisen G. Preterm delivery and calculation of sur-vival rate below 28 weeks of gestation. Acta Paediatr 2003;92:1335e8.

3. Larroque B, Breart G, Kaminski M, et al. Survival of very pre-term infants: Epipage, a population based cohort study. ArchDis Child Fetal Neonatal Ed 2004;89:F139e44.

4. Field D, Petersen S, Clarke M, Draper ES. Extreme prematurityin the UK and Denmark: population differences in viability.Arch Dis Child Fetal Neonatal Ed 2002;87:172e5.

5. Costeloe K, Hennessy E, Gibson AT, Marlow N, Wilkinson AR.EPICure Study Group. The EPICure Study: outcomes to dischargefrom hospital for infants born at the threshold of viability. Pediat-rics 2000;106:659e71.

6. Marlow N, Wolke D, Bracewell MA, Samara M. EPICure StudyGroup. Neurologic and developmental disability at six yearsof age after extremely preterm birth. N Engl J Med 2005;352:9e19.

7. Nuffield Council on Bioethics. Dilemmas in current practice:babies born at the borderline of viability. Critical Care Deci-sions in Fetal and Neonatal Medicine: Ethical Issues. November2006, p. 67e87.

8. Tommiska V, Heinonen K, Lehtonen L, et al. No improvement inoutcome of nationwide low birth weight infant populations be-tween 1996e1997 and 1999e2000. Pediatrics 2007;119:29e36.

9. Marlow N. Outcome following extremely preterm birth. CurrObstet Gynecol 2006;16:141e6.

� Whereas the definition of ‘viability’ is open to de-bate, there are no ‘margins of vitality’. An infanteither has signs of life or has not e however un-comfortable that might be.� The NICU should be seen as providing a supportive

role to parents while their baby is essentially un-dergoing a ‘trial of life’. Decisions on withdrawalof care should take into account not only the per-ceived futility of treatments, but also the mount-ing burden of neonatal complications.

Research directions

� The use of neonatal networks to share informationon treatments and outcomes of babies of border-line viability.� Large multicentre RCTs to determine the benefits

of traditional fetal and neonatal treatments inpopulations of extremely low gestation.� The use of minimally invasive resuscitation and in-

tensive care for infants of borderline viability.� The role of antenatal factors in the prediction of

outcomes.� The role of neonatal imaging and neurophysiologi-

cal investigations in the prediction of neurodisabil-ity in babies of borderline viability.

Infants of borderline viability 15

10. Lagercrantz H. The emergence of the mind e a borderline ofhuman viability. Acta Paediatr 2007;96:327e8.

11. Annas GJ. Extremely preterm birth and parental authority torefuse treatment e the case of Sidney Miller. N Eng J Med2004;351:2118e23.

12. Iams JD, Mercer BM. National Institute of Child Health andHuman Development MaternaleFetal Medicine Units Network.What we have learned about antenatal prediction of neonatalmorbidity and mortality. Semin Perinatol 2003;27:247e52.

13. Payot A, Gendron S, Lefebvre F, Doucet H. Deciding to resus-citate extremely premature babies: how do parents andneonatologists engage in the decision? Soc Sci Med 2007;64:1487e500.

14. Bastek TK, Richardson DK, Zupancic JAF, Burns JP. Prenatalconsultation practices at the border of viability: a regional sur-vey. Pediatrics 2005;116:407e13.

15. Keenan HT, Doron MW, Seyda BA. Comparison of mothers’ andcounselors’ perceptions of predelivery counselling for ex-tremely premature infants. Pediatrics 2005;116:104e11.

16. Greisen G. Managing births at the limit of viability: the Danishexperience. Semin Fetal Neonatal Med 2004;9:453e7.

17. Hansen BM, Hoff B, Greisen G, Mortensen ELDanish ETFOLstudy group. Early nasal continuous positive airway pressurein a cohort of the smallest infants in Denmark: neurodevelop-mental outcome at five years of age. Acta Paediatr 2004;93:190e5.