Article Text

Download PDFPDF

Withdrawal of artificial nutrition and hydration in the Neonatal Intensive Care Unit: parental perspectives
  1. Jonathan Hellmann1,
  2. Constance Williams2,
  3. Lori Ives-Baine3,
  4. Prakesh S Shah2
  1. 1Division of Neonatology, Hospital for Sick Children, Department of Paediatrics, University of Toronto, Toronto, Canada
  2. 2Department of Pediatrics, Mount Sinai Hospital, Toronto, Canada
  3. 3Neonatology Program, Hospital for Sick Children, Toronto, Canada
  1. Correspondence to Jonathan Hellmann, Division of Neonatology, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada; jonathan.hellmann{at}


Objective To review the experience of the practice of withdrawal of artificial nutrition and hydration (WANH) and to describe parental perspectives on the process.

Design A retrospective chart review and parental survey.

Setting Tertiary level Neonatal Intensive Care Unit (NICU).

Participants Infants who had WANH after withdrawal of other life-sustaining treatment, and their parents.

Main outcome measure Parental perspectives on the care and process were obtained through a survey administered 1 to 4 years after the death of their infant.

Results Fifteen cases (5.5% of all mortality and 0.5% of all admissions) of WANH were identified, and 10 parents participated in the survey. The median (range) gestational age was 40 weeks (31–42) and birth weight was 3409 g (2000–4640). The reason for WANH was predicted poor outcome due to severe neurological injury/disease. The median (range) time between WANH and death was 16 days (2–37). All parents reported favourable perceptions of preparation, support, communication and care. Seven parents reported concerns regarding pain experienced by their infant. Parents reported the ability to spend quality time, creating tangible memories and the virtues and professional qualities of the caregivers to be helpful, but identified that consistency and continuity of care could be improved.

Conclusion Within the spectrum of palliative care in neonates, WANH can be a tenable, justifiable and humane practice in the NICU.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic

  • Controversy exists when discussions or decisions regarding withdrawal of hydration and nutrition take place in the setting of end-of-life care.

  • Parental perspectives from those who have undergone this process are yet to be determined.

What this study adds

  • Withdrawal of artificial hydration and nutrition in neonates is a difficult but possible and humane practice.

  • Parents reported satisfaction with the process; however, they also indicated areas for improvement.


Neonatal caregivers face a dilemma when the initiation or continuation of active treatment of a critically ill newborn may result in a prolonged dying process or the survival of an infant with an extremely poor neurodevelopmental outcome. Palliative care, where a cure-oriented approach is replaced by a spectrum of management options to prevent and relieve pain and improve the transition to dying, is slowly being integrated into practice in the Neonatal Intensive Care Unit (NICU). The slow adoption of palliative practice may be due to a number of factors: uncertainty in determining prognosis; difficulty in moving away from an interventionist approach when therapeutic options still seem possible; divergent caregiver perceptions of the ‘right’ action; moral and legal ambiguities; fear of a ‘lingering death’1; the tendency for the compartmentalisation of medical care into specialist teams and a lack of formal training and experience in palliative aspects of neonatal care.2

Until fairly recently, withdrawal of life-sustaining treatment (WLST) consisted predominantly of withdrawal of assisted ventilation and withdrawal or withholding of pharmacological support for the cardiovascular system, usually resulting in death within minutes to hours. However, when critically ill newborns ‘linger’ after WLST, with continued neurological impairment and poor long-term outcome, the withdrawal of artificial nutrition and hydration (WANH) may become a consideration. Although the practice is emerging as an accepted part of palliative care in adults, WANH is a significant challenge for practitioners in neonatal medicine.3 Recently, the American Academy of Pediatrics and the Canadian Pediatrics Society released updated practice points on the subject of WANH in children. Both identify the controversies facing practitioners and acknowledge the legal and ethical permissibility of WANH. Both highlight the critical importance of the role of parents in shared decision-making, and suggest particular attention be paid to the individual context of each case and the psychosocial needs of parents and families.4 ,5 Other studies have explored the experience of parents following WLST in the newborn and paediatric populations,6,,9 but none has focused specifically on the withdrawal of artificial nutrition and hydration. The objective of this study was to review the practice of WANH in one NICU, and to describe the perspective of parents of infants in whom this approach has been undertaken.



This study was conducted in the NICU at the Hospital for Sick Children in Toronto, Canada. All neonatal deaths between January 2004 and August 2007 were reviewed. Patients were eligible for inclusion if there was a specific decision for WANH following withdrawal of mechanical ventilator support. We excluded those patients in whom withdrawal of respiratory support and cessation of feeding were undertaken at the same time and death occurred within the subsequent 24 h, as this did not constitute a conscious elective decision regarding WANH.


Over the 4 years of the review period, the practice of WANH was a formalised process as described below. This was developed after a consultative process with the healthcare team members and ethicist. No approval from the research ethics board was sought for the process as it was considered part of routine care. In the majority of cases, WANH was presented to the parents as an option of care if the attending team determined it to be in the infant's best interest. In a few cases parents raised the idea themselves. Decisions always involved the inter-disciplinary team as well as discussions with an ethics consultant, or coroner as appropriate, or both. The coroner's role was for consultation regarding after death process and not in the decision-making. The process involved the following, with the wishes of the parents foremost in decision-making: (1) identification of the prognosis for each infant as accurately as possible; (2) the clear demonstration of the inability of an infant to safely tolerate oral feeds (oral feeding was regarded as routine care even in those in whom it took many hours to achieve a satisfactory intake); (3) the full agreement of the parents to pursue such a course for their infant; and (4) the full agreement of the senior medical staff and the team directly involved in the care of the infant.


Care of each infant took place in a private ‘care-by-parent room’ with support of the family by core nursing and medical teams, as well as the palliative care and bereavement coordinator. An accepted principle in the NICU is that staff with moral misgivings may opt not to participate in this portion of end-of-life care (no staff member over the review period chose to opt out of this aspect of care). A care plan with respect to comfort measures including the use of non-pharmacologic and pharmacologic interventions for analgesia and sedation was developed for each infant. All forms of artificial nutrition and hydration (ANH) were withdrawn at the same time in all infants.

Components of this study

Part 1: retrospective chart review

A chart review was undertaken to collect the demographic details, diagnoses, the time to end-of-life decision-making, time from the end-of-life decision to the implementation of WANH, details of the interaction with parents from various members of the healthcare team, the medical management and various timelines. In addition, qualitative aspects of care, such as nursing comments regarding the infant's condition, and documented comments regarding parental perspectives were collected.

Part 2: parental perspectives

In order to obtain the parents' perspectives on the practice and care received, parents were asked to complete a brief questionnaire mailed to them after a minimum of 1 year following the death of their child. This time lapse was chosen to minimise the potential for undue parental distress. In addition to five questions with a Likert scale10 range of responses (1–5), parents were given the opportunity to add open-ended comments. The questions addressed how well the parents were prepared for changes in their baby, what support they received from the NICU team, the frequency of meetings with staff, what pain or discomfort they felt their baby experienced, the response of the team to their infant's or their own concerns and the overall quality of end-of-life care they, and their child, received. Finally, parents were asked to comment on what was done well and what improvements could be suggested, and their reaction to the survey and participation in the study. No personal interviews were conducted as it was felt that reliving details of their child's death in front of an interviewer might distress some parents, and thus only parents' written responses were collected and analysed.


The data from the chart review were analysed using simple statistics of median and range. The qualitative aspects of the study were analysed independently by two members of the study team. Responses underwent concept saturation (until no new concepts arose from the responses) and theme generation, and were sorted into categories and subcategories. The categories were revised via a process of reflection and discussion by two study authors. Trustworthiness of emerging themes was verified via reference to quotes in the text. Negative case analysis was performed in which exceptions to the emerging themes were specifically sought and investigators discussed the ‘fit’ of the emerging findings. With the limited sample size, the responses were categorised one-by-one as they were received, to gauge concept saturation.

Ethical approval

Approval was obtained from the Research Ethics Board at the Hospital for Sick Children, Toronto, Canada. In view of the sensitivity of the study, parents were first asked permission for researchers to review their child's medical record, and thereafter invited to receive a mailed questionnaire to provide feedback on their experience. Parents were able to choose to participate only in the retrospective chart-review phase of the study and not complete the questionnaire.


Part 1: chart review

During the study period, of a total 2741 admissions, 271 neonates died (10% of total admissions). Withdrawal of artificial hydration and nutrition was undertaken in a total of 15 infants (5.5% of all deaths and 0.5% of all admissions). Of these 15 infants, three parents had changed location, and one parent did not respond to mailed letters. Of the 11 sets of parents who consented, nine infants had severe hypoxic-ischaemic encephalopathy, and two infants had other causes of severe neurological injury (the details are not given to preserve confidentiality of the infants due to the rarity of the disorders). During their hospital course, all neonates had been ventilated, received anticonvulsants (before WANH) and intravenous fluids, before discussions of withdrawal of assisted ventilation (WAV). The reason for WAV in all infants was based on the prediction of a poor outcome due to the devastating neurological injury or disease. None of the neonates could be safely orally fed. The median time between WAV and WANH was 2.5 days (1–13 days). All neonates received routine mouth and skin care as well as analgesia. Nine infants received sedation after WANH. Most infants died in hospital and two died at home with their parents. In six of the 11 cases, the parents consented to an autopsy (twice the average rate for our unit) which confirmed the clinical findings in all cases. Demographic details and timelines are described in table 1. Because of privacy and confidentiality purposes, we have not provided a detailed table of the characteristics of the individual patients.

Table 1

Demographic details of study patients (n=11)

Part 2: parental perspectives

Of 11 parents who consented to the chart review, 10 completed the questionnaire. We did not specifically ask which parent was involved in the completion. Parental responses showed that they felt well prepared and supported by the neonatal and palliative care and bereavement team. Though they were concerned about the amount of pain experienced by their infant, they indicated that their infant's symptoms were adequately managed and that the infants received good quality end-of-life care (table 2). Appreciation for the ability to provide feedback to the care teams through this study was highlighted by several parents. Two parents wrote: ‘it brought back some of what happened and made me sad but I will be glad if his story helps make a difference’ and ‘it is admirable and a testament to the hospital’s commitment to excellence and growth that you are looking for feedback and constructive criticism'.

Table 2

Parental responses to the withdrawal of artificial nutrition and hydration process (n=10)

From the open-ended questions illustrating the positive and negative aspects of parent experiences, three themes emerged through the qualitative analysis: communication, decision-making and the dying experience. These are summarised, with quotes and examples in figure 1.

Figure 1

Themes, subthemes and responses of parents regarding the withdrawal of artificial nutrition and hydration process.

The subthemes which emerged in the communication domain included the virtues of good communication, the importance of relationship-building, admitting the uncertainty of prognoses, and ways of sharing information. The special relationship between caregivers and family is described by this parent: ‘I told the nurses to take pictures for us when we couldn’t be there … to snuggle with (her) as much as possible. She was a very popular girl. I would call in to see how she was doing and they would tell me that she was on rounds with a nurse … Nurses would stay after their shift to snuggle with her. I even called in one day and the lady at the front desk (ward clerk) had her.' Parents also pointed out some of the relationship challenges. One parent described when team consistency breaks down: ‘The team is constantly changing, so few people actually know the patient and can recognise how the child is doing/changing over time. Consistency of care and continuity will provide parents with increased peace of mind and will help reduce overall stress and anxiety.’

The subthemes in the ‘Decision-making’ domain included decision-maker reluctance, the decision itself and the process. The parents described the extreme difficulties they encountered as decision-makers: ‘Deciding whether to take your child off food is the worst thing you will ever do. It rips your heart out again and again … Keeping her would have been only for us. She would never get out of bed … It would be like having a living doll around and that’s not right … The decision killed us but we felt like it was our only choice. The only thing that made it just a little easier is the fact that she didn't have the brain function to know she was hungry. But we knew!'

The final theme, the ‘Dying experience’ had the following subthemes: the myriad of parent emotions, making the best in a bad situation, receiving encouragement and support, and making a life after death. The median time between WANH and death in our study was 16 days. One parent described the ambivalence of waiting for their child to die: ‘They took all the IVs out … the next 2 weeks were heaven and hell. We would feel great with her. The minute we would leave we would feel like …. If the phone would ring we would jump out of our skin thinking it was the hospital. In a way we wanted that phone call and then again we didn’t.' In their own words, parents express their raw emotions; their feelings of anger and guilt and their search for forgiveness: ‘I know in my brain that it wasn’t my fault but telling my heart is a different story. Feeling like you have killed your daughter twice makes you feel … I can't tell you. We have asked for forgiveness from (her) for what we have done to her. We will never know that answer and the guilt we will have to live with for the rest of our lives.' This mother's anguish is palpable in her words. She goes on, however, to write beautifully of living, of healing and finally of finding hope with a message to her daughter: ‘(She) lived 29 days and in that time she showed us more than we could ever dream of. Fly as high as you can!’


Public acknowledgement of the presence of WANH in the modern era of neonatal intensive care is minimal. The first published case in which this issue was raised was in 1988 when the parents of a full-term infant with severe birth asphyxia asked the healthcare providers to consider WANH so as to allow their son to die.3 The healthcare workers were reluctant and ultimately parents chose to continue comfort measures and to provide food and water via a nasogastric tube. Since that time, other cases have illustrated the challenges in WANH in the NICU.11,,13 These raise issues of uncertainty, the wish to prevent and relieve pain, moral distress of staff, and stories of lingering death with withdrawal of nutrition without withdrawal of hydration.

A frequent debate surrounds the question of whether ANH is a medical treatment or an obligatory part of simple humane care.3 ,14 All forms of maintaining ANH (the passage of nasogastric tubes, the insertion of intravenous needles or the placement of a gastrostomy tube) are invasive to a varying extent, and carry medical risks of dislodgement, migration, error, infiltration and infection. There appears to be general agreement in the adult literature, at least over the past two decades, that artificial hydration and nutrition is a medical treatment, at par with mechanical ventilation and other life-sustaining technologies.15 There is support for the view that decisions about hydration and nutrition should be consistent with the underlying condition, prognosis and goals of care and should not be held to a higher standard.16

The ethical justification for WANH in adults with severe, unrelieved pain and advanced incurable illness is based on the concept of honouring a competent patient's right to refuse such treatment. Justifying WANH is an individual act as opposed to a policy17 and requires each unique context to be explored, with full parental understanding of consequences, assurances of comfort measures and with appropriate cultural and spiritual sensitivity.18 The issue in neonates is whether parents, as surrogate decision-makers, should be accorded these same rights to make decisions which they and the medical team feel to be in the best interest of the child. Children are different, but this in of itself does not preclude WANH if deep-seated psychological factors in the symbolism of feeding can be assuaged, and pain can be diminished as much as possible. To our knowledge, there are no reports on the successful integration of WANH into NICU practice or parental perceptions of the process. This study presents the first examination and report on the process of WANH in the neonatal population. Importantly, it incorporates a description of the process and feedback from parents, the surrogate decision-makers for patients undergoing this medically and ethically challenging aspect of palliative care.

In our study, parents ranked the overall quality of end-of-life care during WANH as high. Items including preparation for the infant's death, degree of support from healthcare workers and communication regarding progress and condition were ranked favourably. In their qualitative comments, they highlighted the positive aspects of care, including the virtues of information-sharing, the ability to build relationships with caregivers, and honesty regarding prognoses and timelines. However, parents also identified the challenging aspects of the process, including themes previously described in the end-of-life literature, the importance of consistency of caregivers and care-plans,19 worry over infant pain and the tragic nature of their situations. Overall, our results support the findings of Catlin, that when the burden of care for a neonate is higher than the benefit, parents need to be our partners in transitioning from curative efforts to pure palliative care.20

During our study period, no healthcare providers opted to excuse themselves from the care of the infants undergoing WANH. We cannot ignore, however, the views of caregivers who regard feeding as part of caring, in the same way as bundling, bathing and holding, even though these views may be of symbolic importance, largely psychological and not ethically or legally compelling. These psychological barriers or ‘traps’ as Leuthner and Carter describe them can be significant stumbling blocks and need to be explored.21 The exploration revolves largely around the concept of starving and the issue of time. Use of the word starving implies pain which does have moral weight and must be treated, but this is no different than the pain of the infant from whom the ventilator is withdrawn. For seven parents in our study, understandably, a concern was the degree to which they felt their infant had pain. These parents appreciated, however, the promptness with which the medical and nursing team responded to signs of pain. Studies in adult patients have shown that death following WANH is from dehydration, not starvation, and that dehydration leads to a decrease in nausea, vomiting, diarrhoea and urine output, with little, if any, discomfort, perhaps because of the release of endogenous opioids with fasting and ketosis.15 ,16 In addition, adult patients experience little, if any, hunger or thirst if appropriate mouth care is provided.3 Jacobs has described this voluntary dehydration as a ‘surprisingly peaceful way to die’.22 Similar to the concept of ‘starving’, time or the fear of prolonged dying as a prohibition against WANH is unlikely to be a valid moral argument, as long as the time is spent without pain and with optimal comfort.19

Finally, the limitations of our study are that this is a small sample-sized retrospective survey of parents who had accepted WANH for their infants. In some cases, the views were expressed 4 years after the death of their child. Future studies may involve inviting parents to share their thoughts and experiences in ‘real time’; however, we recognise the complexity of designing a sensitive, prospective study of this nature.


WANH is ethically permissible once a decision is made to pursue a palliative approach to care, and feeding is perceived as a burdensome intervention that is not in the best interest of the infant. Nevertheless, the great symbolism of feeding newborns and the psychological impact of WANH on parents and the healthcare team requires moving slowly and with caution. Full transparency, morally contoured communication and support during decision-making, throughout the dying process, and after the death of the infant, enable the practice of WANH to be performed in a compassionate way. Thus, we believe that in specific cases, WANH can be a tenable, justifiable, and most importantly, a humane practice in the NICU.


The authors would like to sincerely thank the parents who have taken time to reflect back on a tragic event in their lives. The authors also express our sincerest gratitude to the multidisciplinary care team who provided such exemplary care to these neonates.



  • Contributors Jonathan Hellmann initiated the concept, wrote the protocol, planned and analysed data and wrote the first draft and critical revision of this manuscript. Constance Williams was involved in the concept, performed chart review, analysed qualitative data and provided input in critical revision of this manuscript. Lori Ives-Baine was involved in the design of the study, collected data, contacted families and provided backup support to families and critically revised this manuscript. Prakesh Shah was involved in the design, protocol, analysis of data and critical revision of the final document.

  • Competing interests None.

  • Ethics approval Research Ethics Board at the Hospital for Sick Children, Toronto.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Additional data are available by contacting Dr Jonathan Hellmann (corresponding author).

Linked Articles