Introduction

Technological and pharmacological advances over recent decades have resulted in improved neonatal outcomes. Neonatologists can now treat infants who previously may not have been resuscitated [25]. However, full resuscitative measures are not always appropriate as they can result in prolonging deaths that are inevitable and therefore prolong unnecessary suffering and limit the quality of the short time that parents have with their children [21, 9].

Deaths within the newborn period generally follow one of four courses: due to life-sustaining treatment (LST) being (1) withheld or (2) withdrawn and (3) occurring despite full intensive care measures or (4) active euthanasia. Studies from across Europe, USA and Australia have shown that newborns are more likely to die following LST either being withheld or withdrawn as opposed to following full resuscitative measures [23, 2, 1, 14, 24, 20, 19, 28, 8, 26]. However, rates of withholding or withdrawing LST vary between countries, and there are many potential reasons to explain these differences, including cultural or religious differences [23, 5, 11, 10]. There have been few reports concerning mode of dying from countries with low rates of termination of pregnancy, such as Ireland, where congenital malformations may account for a relatively greater proportion of neonatal deaths.

Our primary objective was to assess modes of dying within an Irish tertiary maternity centre. We also reviewed causes of death, location of death within the hospital, decision-making processes involved and use of analgesia in end-of-life care in the newborn infant in an Irish setting.

Methods

This was a retrospective chart review of neonatal deaths over a 36-month period in an Irish tertiary maternity centre. There are approximately 8,800 deliveries each year in Cork University Maternity Hospital, with approximately 1,200 admissions annually to its neonatal intensive care unit. All patient records of newborn deaths within our hospital between January 2010 and 2013 were reviewed. We recorded relevant demographics including sex, gestational age, birth weight and mode of delivery.

Deaths were categorized into one of four categories according to the Wigglesworth classification system: congenital malformation, prematurity, hypoxic ischemic encephalopathy or specific [27, 16]. The mode of dying was classified as has been described previously, but active euthanasia is not legal and is not recorded in the Republic of Ireland [6]. Therefore, deaths occurred following withholding or withdrawal of LST or despite of maximal intensive care efforts. Deaths that followed withdrawal of LST were subdivided into two separate categories based on either physiological instability or quality of life determinations [23]. Infants who died despite ongoing maximal intensive care efforts died on a ventilator. Those infants who died following the withdrawal of LST in a moribound state were extubated and spent time in their parents’ arms, although for a short period of time. Modes of dying were therefore classified as:

  1. 1.

    Withholding LST: death attributable to withholding of treatment necessary for immediate survival including surgical intervention, resuscitation measures, respiratory support and administration of adrenaline.

  2. 2.

    Withdrawal of LST in moribund infants: death attributable to elective discontinuation of ongoing life support in infants who were deemed to be physiologically unstable. To be classified as unstable, infants had to have two of the following criteria: persistent desaturation despite 100 % oxygen on mechanical ventilation, hypotension despite volume infusion and inotropes, protracted bradycardia or anuria for >24 h.

  3. 3.

    Withdrawal of LST for quality-of-life reasons: death due to elective discontinuation of ongoing life support in physiologically stable infants who did not meet the above criteria.

  4. 4.

    Death despite ongoing maximal intensive care treatments.

The timing of decisions and whether the parents were involved in this process was assessed. Timing of death following decisions to withdraw or withhold LST and use of palliative medicines during this time was recorded.

Results

There were 64 neonatal deaths in our hospital during the study period from approximately 26,500 deliveries, an overall neonatal mortality rate of approximately 2.4/1,000 births. The mean gestational age was 32 weeks and the mean birth weight was 1,875 g. Multiple births accounted for 9 % (n = 6) of neonatal deaths. Just over half of the newborns (n = 33, 52 %) were delivered vaginally. The neonatal unit was the most common place for newborns to die within our hospital (n = 38, 59 %) followed by the delivery room (DR) (n = 21, 33 %). The principal reasons for death in the delivery room were attributable to antenatally diagnosed underlying chromosomal disorders and periviable maturity.

The categories to which neonatal deaths were attributed to are displayed in Fig. 1. Congenital abnormalities were the most frequent cause of neonatal deaths, accounting for 47 % (n = 30) of all deaths. Chromosomal abnormalities were detected in 37 % (n = 11) of those deaths due to congenital abnormalities. They consisted of six infants with trisomy 18, two infants with triploidy and single cases of trisomy 13, polycystic kidney disease and Smith Lemli Opitz syndrome. The most frequent causes of non-chromosomal disorders were cardiac conditions incompatible with life (n = 5), skeletal dysplasia (n = 4) and pulmonary hypoplasia (n = 4).

Fig. 1
figure 1

Deaths by category

Full size image

Prematurity and complications arising from prematurity was the most common cause of death in the neonatal intensive care unit (n = 19, 50 %). The reasons included progressive respiratory failure (n = 10), sepsis (n = 6) and perforated necrotising enterocolitis (n = 3). The specific group comprised of one death due to a cervical haemorrhage and another following small intestinal bowel ischemia secondary to a gastrointestinal volvulus associated with congenital malrotation.

Neonatal deaths are displayed in Fig. 2 by mode of death. Withholding LST was the most frequent mode of dying and occurred in 38 % (n = 24) of all deaths. Congenital malformations (n = 18) accounted for 75 % of cases where LST was withheld. Life-sustaining treatment was withdrawn in 50 % of neonatal deaths (n = 32). Median (range min–max) time for survival post-extubation was 28 (4.8–120) min in those extubated in a moribund state and 5 (0.5–428) h in those extubated for quality-of-life reasons. One infant born prematurely was initially resuscitated in the DR but had LST withdrawn for quality-of-life reasons on the arrival of a senior neonatologist. All infants with congenital abnormalities that had LST withheld, who subsequently died in either the DR (n = 13) or PNW (n = 5), had an antenatal diagnosis with the decision to withhold LST made prior to delivery.

Fig. 2
figure 2

Mode of death

Full size image

Parental involvement in end-of-life decision-making was documented in 88 % (n = 56) of the charts. In the remaining 12 % (n = 8) of cases, death was unexpected and occurred despite maximal therapy. Where LST was withheld or withdrawn, 32 % (18/56) of decisions were made antenatally, and in a further 25 % (15/56) decisions were made within the first day of life.

Morphine was the analgesic of choice for palliation within the neonatal unit. A total of 94 % (29/31) of infants who had LST withheld or withdrawn in the neonatal unit had intravenous morphine as palliation. The two infants who did not receive palliative medical treatment were extubated in a moribund state and died in their parents’ arms within 15 min of LST being withdrawn. Oral sucrose was utilized in the delivery suite if clinically deemed necessary, and oral sucrose and small comfort feeds were utilized for deaths that occurred on the postnatal ward. The administration of opiates outside the setting of the intensive care unit was not routine.

Discussion

This study provides an insight into end-of-life care within an Irish maternity centre. The main findings were that congenital abnormalities are the most frequent cause of death. Death following life-sustaining treatment being withheld was the most frequent mode of death, and deaths occurred in a patient-centered environment where parents were involved in the majority of end-of-life decisions.

Congenital abnormalities were the most frequent cause of death, accounting for approximately one half (47 %) of overall deaths in our centre. Prior Irish studies have also found that congenital malformations account for a significant proportion of neonatal deaths [13]. This finding is not surprising as, for both legal and cultural reasons, rates of medical terminations are low amongst our population. In Europe, approximately 17 % of all pregnancies where congenital abnormalities are identified result in medical termination [7]. In our maternity centre, we do not record when pregnancies with known congenital anomalies are medically terminated abroad; therefore, true rates are not known. Infants with trisomy 18 accounted for almost 10 % of our newborn deaths. The prevalence of trisomy 18 is similar in countries across Europe, but rates of termination differ dramatically (reported rate of 7 % in Ireland compared with a European average of over 70 %) [17].

End-of-life practice within maternity centres has been shown to differ both within and between countries. In general, the pattern that has emerged over the past 20 years has shown that increasing numbers of newborn deaths follow either LST being withheld or withdrawn [23, 2, 1, 14, 24, 20, 19, 28]. However, this trend is not universal. Recent studies from Israel and Latin America displayed high numbers of infant deaths following full resuscitative measures [10, 11]. This is in stark contrast to findings from The Netherlands, where virtually all newborn deaths follow the withdrawal of LST [22]. Cultural differences are thought to play a role, but this does not explain why differences persist in maternity centres within the same jurisdiction. Newborn deaths following full resuscitative measures in two North American centres within the same study were shown to differ dramatically (31 versus 16 %) [23]. Within our study, 12 % of deaths occurred during full resuscitative measures, which is comparable to findings from other North American and European centres [23, 2, 1, 14, 24, 20, 19, 28].

In Ireland, it was previously reported that 77 % of paediatricians had either withheld or withdrawn LST in babies thought to have a hopeless outcome [12]. Within our study, 88 % of newborn deaths followed LST being withheld or withdrawn. Studies from North America, UK, Australia and across Europe have reported similar majorities ranging between 58 and 93 % [23, 2, 1, 14, 24, 20, 19, 28]. Of those infants who died following LST either being withheld or withdrawn, a higher proportion of deaths in our study followed LST being withheld (38 %) compared with other centres (0–24 %). We believe that two factors are responsible for this difference. Firstly, pregnancies with known lethal congenital malformations are less likely to have resulted in termination within our population and so progress to delivery. For these infants, palliation is commenced at birth in the delivery suite. This is very much dependent on the individual circumstances of the case and the wishes of the parents. Secondly, our centre does not have pediatric surgical facilities on site, and therefore a number of infants born in our centre who ultimately have LST withdrawn may die elsewhere. This may have resulted in a lower rate of death following withdrawal of LST being reported in our study than what actually occurs among our population, although in exceptional circumstances where intervention is considered not to be in the patient’s best interest, the patient may return for palliative care in our centre.

End-of-life care should involve an interdisciplinary approach from the time of initial diagnosis to the time of a comfortable and dignified death [3, 4]. In our centre, we believe that all end-of-life decisions should be a shared process between obstetrical staff practising foetal medicine, neonatal staff and parents. This was reflected in our study as parents were involved in all decisions in which death followed LST being withheld or withdrawn. All infants who died within our study despite ongoing maximal intensive care measures had an unexpected clinical deterioration at the time of their death (for example, a large pulmonary haemorrhage), resulting in inadequate time to counsel or reach fully informed assent from parents regarding withdrawal of LST or to initiate palliative measures. These infants died on a ventilator with ongoing CPR.

In all cases where withdrawal or withholding of LST occurred, a joint decision was made without the need for recourse to ethics committees or legal intervention, which we believe is a reflection of the effective communication process that takes place. All parents of infants with a confirmed antenatal diagnosis of a lethal congenital abnormality, where the prognosis is guarded, meet with a neonatologist, obstetrician in foetal medicine and a bereavement counsellor prior to delivery. A detailed discussion takes place and a clear management plan is prepared and documented. This addresses issues such as presence of senior personnel at delivery to direct the process, ensuring family wishes such as the presence of other family members or in some cases a member of the clergy should the family wish to have their newborn blessed, oral sucrose for comfort and contingency plans in place should the newborn survive the first few hours of life. For a small number of infants, life expectancy exceeds prenatal expectations. Such infants with guarded prognoses may live for days or even months. This clinical uncertainty is discussed and acknowledged with parents at antenatal consultation. In such cases, further multidisciplinary meetings take place postnatally where revised palliation plans are agreed upon with the parents, and community-based palliative supports are arranged to facilitate discharge of infant home with family.

One third of all newborn deaths occurred within the delivery room or postnatal ward, which displays our reluctance to separate parents from their dying children and to ensure that privacy and family involvement is central. Of those admitted to the neonatal unit, almost all infants had intravenous opioid medications as comfort care. This is higher than previous reports in Ireland and similar to findings in international studies [15]. This higher rate of opioid administration is recognition that newborn deaths may be associated with discomfort for the newborn, and the aim of the physician is to relieve this discomfort. However, outside of the neonatal unit, the use of palliative medicines was limited to the administration of oral sucrose, but the dose was both variable and inconsistent. The administration of opioids outside the setting of the neonatal intensive care unit is not routine in our service, and the majority of babies who pass away in the delivery suite do so peacefully, and relatively quickly, in their parent’s arms. We recognize that this is an aspect of care that requires further evaluation. National guidelines providing a structured approach to medical palliation in the newborn infant may improve this aspect of care as seen in other countries [18].

In conclusion, our study further establishes that newborn deaths occur differently in maternity centres from culturally different backgrounds. There is a higher rate of LST being withheld within our centre than in other countries, which we feel is at least partially due to the legal status and cultural attitudes toward medical termination of pregnancy in Ireland.