Article Text
Abstract
Objective Many extremely preterm neonates die in the delivery room (DR) after decisions to withhold or withdraw life-sustaining treatments or after failed resuscitation. Specific palliative care is then recommended but sparse data exist about the actual management of these dying babies. The objective of this study was to describe the clinical course and management of neonates born between 22 and 26 weeks of gestation who died in the DR in France.
Design, setting, patients Prospective study including neonates, who were liveborn between 22+0 and 26+6 weeks of gestation and died in the DR in 2011, among infants included in the EPIPAGE-2 study at the 18 centres participating in this substudy of extremely preterm neonates. Data were collected by a questionnaire completed by the professional caring for each baby.
Results The study included 73 children, with a median (IQR) gestational age of 24 (23–24) weeks. Median (IQR) duration of life was 53 (20–82) min. All but one were both wrapped and warmed. Pain was assessed for 72%, although without using any scale. Gasping was described for 66%. Comfort medications were administered to 35 children (50%), significantly more frequently to babies with gasping (p=0.001). Mother–child contact was reported for 78%, and psychological support offered to parents of 92%.
Conclusions Non-pharmacological comfort care and parental support were routinely given. Comfort medication was given much more frequently than previously reported in other DRs. These data should encourage work on the indications for comfort medication and the interpretation of gasping.
- extreme prematurity
- comfort care
- delivery room
- Palliative Care
- dying
Statistics from Altmetric.com
What is already known on this topic?
When resuscitation is not performed or is unsuccessful, comfort care is recommended for extremely preterm neonates dying in the delivery room.
These babies who die in the delivery room with comfort care receive medication less often than similar babies dying in the neonatal intensive care unit.
What this study adds?
Nearly all the newborns received non-pharmacological comfort care, and most parents were offered support, mementos and contact with their baby.
Half the infants received comfort medication, much more than in previous delivery room reports. Medication was given most frequently to babies who were gasping.
These results underscore the difficulty of interpreting gasping in this context since it is usually not associated with suffering and medication routes are invasive.
Introduction
The survival of extremely preterm neonates has increased in the past 20 years, and curative management is offered in many countries. Population-based studies show high mortality for children born between 22 and 26 weeks,1–3 and this rate depends partly on hospital practices for the initiation of active care (ie, life-saving, life-sustaining. or intensive-care procedures).4 A French population-based study, the EPIPAGE-2 study conducted in 2011, reported survival to discharge below 1% for infants born alive before 24 weeks of gestational age (GA), 31% for those born at 24 weeks, 59% at 25 weeks and 75% at 26 weeks.3 ,5 In many countries, these deaths follow a decision to withhold or withdraw active care, according to guidelines based on GA, individual assessment by physicians and parental wishes: no resuscitation at birth at 22 weeks, active care at 25 weeks and more variable guidelines for infants born at 23 and 24 weeks.6–8 Various groups had begun describing specific principles of comfort care for infants dying in the delivery room (DR) by 20119–15: warming and wrapping the baby, minimising pain/suffering and environmental aggression, family support including offering contact with the baby, the creation of mementos and support by professionals trained to deal with perinatal grief. In France, guidelines on management of extremely low GA neonates (ELGANs) published in 2010 by a perinatal group only briefly mentioned the nature of ‘palliative care’, without detailed recommendations.8 Since then, many countries have published guidelines recommending comfort (or palliative) care for ELGANs not receiving active care immediately after birth.7 ,16 Nonetheless, little is known about the clinical course of these babies and especially the procedures actually used for their care: we found four retrospective studies, based on chart reviews.9 ,16–18 Most studies have focused on deaths in neonatal intensive care units (NICUs).17 ,19–23 Our principal objective was to describe the clinical course and management of babies born at 22–26 completed weeks of gestation in France who died in the DR. Our secondary objectives were to describe the offer of support for subsequent parental mourning, the support provided to the staff and their satisfaction about the quality of care provided.
Data and methods
Study population
EPIPAGE-2 (Etude Epidémiologique sur les Petits Ages Gestationnels) is a population-based cohort of preterm children born between 22 and 34 weeks of completed gestation in France from May to December 2011.24 This paper is based on a specific substudy designed to assess the management of extremely preterm infants who died in the DR and included all infants enrolled in the EPIPAGE-2 cohort, born alive between 22 weeks+0 and 26 weeks+6 of gestation in one of the 18 participating maternity units (17 level-3 units of the 65 in the main study and one of the 214 level-2 units). Exclusion criteria were: termination of pregnancy; stillbirth and birth in a non-participating centre. All centres from the EPIPAGE-2 study were asked to volunteer to participate in this substudy; only 18 finally did. They were located in 13 different regions in France. Professionals involved in the writing of the French guidelines for ELGAN management in 20108 worked at 10 of the 18 participating centres. Nine centres had written policies for the care of babies dying in the DR.
Ethics statement
Families received information and agreed to participate in the study before data collection began. The National Data Protection Authority (CNIL no 911009) and the appropriate committees, that is, the Consultative Committee on the Treatment of Information on Personal Health Data for Research Purposes (no 10.626) and the Committee for the Protection of People participating in Biomedical Research (no CPP SC-2873), approved this study.
Data collection
Two sets of data were used for the study. The first came from the EPIPAGE-2 population data, collected as described elsewhere3 and included GA, birth weight, sex, mother’s age and time from birth to death. GA was expressed in completed weeks of gestation and defined as the best obstetrical estimate combining last menstrual period and ultrasound assessment. The second came from this specific substudy of extremely preterm infants. Data were collected with a questionnaire completed by the professional (neonatologist or midwife) involved in the infant's care at the time of his/her death, or within hours or days after the death. Data concerning the circumstances of death included resuscitation, defined as any or all of endotracheal or nasal intermittent positive pressure ventilation, chest compressions, intravenous or intratracheal epinephrine and fluids;25 administration of comfort care including wrapping and warming; medications (yes/no, if yes, pharmacological class or substance and route of administration); pain assessment (yes/no, score used); gasping (yes/no, < or ≥15 min); each parent's contact with the child (seeing, touching or holding); creation of mementos; psychological support to parents in the DR (offer/acceptance); offer of a subsequent consultation with the medical team; psychological support for staff after the death; the staff's satisfaction with the care provided (Likert-type scale ranging from 1=very satisfied to 5=very dissatisfied).
Statistical analysis
Quantitative variables were expressed as medians (IQR) or means (±SD) according to their distribution. Qualitative variables were expressed as n (%). Quantitative variables were compared with the Wilcoxon-Mann-Whitney test, and qualitative variables with the χ2 test. Stata SE V.11.0 (College Station, Texas) was used for all analyses, and p values <0.05 were considered statistically significant.
Results
Study population
The 18 participating centres submitted 419 questionnaires for 533 theoretically eligible infants, that is, born at a GA of 22 weeks+0 to 26 weeks+6 (see figure 1). Only 10 of the 114 eligible but not included infants died in the DR. Of the 302 included children born alive at 22–26 completed weeks, 73 died in the DR (25%). Table 1 summarises their characteristics. The 10 infants not included who died in the DR did not differ significantly from the 73 included for GA (mean (SD): 23.3 (1.2) vs 23.8 (1.0) weeks; p=0.11) or birth weight (mean (SD): 541 (80) vs 620 (128) g; p=0.07). The median (IQR) number of DR deaths per centre was 3.5 (2–6) (range 1–9).
Practices reported
Circumstances of death and clinical management
Of the 71 babies with available information, 53 (75%) were not resuscitated, while 18 (25%) were. Among the latter, 5 (7%) died after unlimited resuscitative measures and 13 (18%) after withdrawal of resuscitation interventions. These included positive pressure ventilation for 16, endotracheal intubation for 9, chest compressions for 4 and intratracheal epinephrine for 2. None received intravenous epinephrine or fluids. Wrapping and warming were reported for 99%.
Symptoms and duration of life
Pain assessment was reported for 72%, although no pain score was used. Gasping was described in 46 (66%), lasting less than 15 min for 26 and longer for 19. The duration of life was recorded for 50 (median (IQR)=51.5 (17–82) min).
Administration of comfort medication
Comfort medication was administered to 35 of 70 infants (50%), by routes summarised in table 2. The type of medication was reported for 21 of the 35. Five infants received two drugs each, and the other 16, only one each. These drugs were opioids (n=14), midazolam (n=8), neuromuscular blockers (NMB) (n=2, always associated with midazolam), phenobarbital (n=1) and propofol (n=1). Neither mean (±SD) birth weight nor median (IQR) duration of life differed significantly between infants who did and did not receive comfort medications (respectively, 634 ±131 vs 605±123 g, p=0.259, and 30.5 (12.5–90.5) vs 56 (20–75) min, p=0.67). Table 3 compares the frequency of comfort medication use according to different categories. Administration of comfort medications varied significantly among GA categories (p=0.02) and was significantly more frequent for babies with gasping (p<0.001). Gasping, prolonged in one case, was reported for both infants receiving NMB.
Contact with mother and/or father, mementos and perinatal grief counselling
Seventy-eight per cent of babies had contact with their mothers, 65% with their fathers and 84% with at least one parent. Mementos were collected in 86% of the cases: photographs in 79%, hand-prints or foot-prints in 29% and others in 26% of the cases. Psychological or psychiatric support was offered to 93% of parents and accepted immediately by 78%. A subsequent appointment with the medical team was offered to 80%.
Staff support and satisfaction with the quality of care
Subsequent psychological support was offered to the perinatal team in 4 of 68 cases (6%). The Likert-like item, completed by 69 providers (95% response rate), showed satisfaction with the quality of the baby's care, with a median (IQR) score of 2 (2–3). Table 4 shows the frequency of satisfied providers according to the characteristics of the situation. Satisfaction was greater when mother–baby contact or father–baby contact occurred (respectively, p=0.03 and 0.03).
Discussion
This study systematically documented the clinical course and management of extremely preterm neonates who died in the DR in 18 units in France. It showed that nearly all the babies received components of neonatal comfort care including wrapping, warming and parental contact. This study thus demonstrated the commitment of DR staff to providing appropriate comfort to babies and their families; it also underscored the absence of support for staff and their moderate satisfaction with the care they provided.
The symptoms and management of dying babies have been extensively described for 20 years, but reports about DR practices are rare.9 ,16–18 Surprisingly, half the babies in our study received medication for end-of-life (eol) comfort. This proportion is high for DR comfort care, compared with other studies, but fairly low compared with NICU deaths. Four studies of DR comfort care have mentioned medication use.16 A study of four units (two in the USA, one in Canada and one in the Netherlands) reported 23 extremely preterm newborns who died with DR comfort care: none received comfort medication;17 a French unit reported that 3 of 19 babies received medication.9 Similarly, a recent study of three units in Germany, excluding babies who died during or after a trial of resuscitation, described 84 preterm neonatal deaths, and only two received such medication. Finally, a study in the Netherlands mentioned that only 1 of 72 liveborn infants who died in the DR received comfort medication.18 In our study, 75% of the babies received primary palliative care, but frequency of medication administration did not differ between primary and secondary palliative care.
Comfort medication is administered to 60–80% of babies dying in NICUs.17–19 Several factors might explain the rarer use of comfort medications in DRs compared with NICUs for eol care. The model of comfort care in the DR aims to avoid unnecessary handling, stress and invasive procedures and to offer babies reassuring skin contact that protects them from pain and anxiety.16 Low gestational and/or postnatal ages are possible reasons to use fewer medications, as found in NICU populations.19 ,20 Finally, the absence of invasive treatments in the DR may minimise the need for comfort medication. Still another reason to use less medication in the DR is simple feasibility and difficulty of administration. In the NICU, infants usually have IV access available at the time of palliative care17 ,19 but not in the DR. The route should be painless and easy to handle to facilitate parental proximity. An umbilical9 ,16 catheter was used in most of our cases, much more frequently than in other studies; a single patient had a peripheral catheter and some had less invasive routes. Little is known about the efficacy of minimally invasive routes; one study has suggested that intranasal fentanyl limits laboured breathing in dying neonates.26
The main question is whether these babies are suffering before death. We know that extremely low GA does not protect babies from feeling pain; indeed, their threshold of pain is particularly low.27 Nonetheless, perinatal analgesia by endogenous substances might protect them from pain.28 In our study, pain/discomfort was assessed subjectively, without the scores normally used in NICUs; as signs of eol discomfort are difficult to assess in these newborns, it is difficult to know when to start or modify treatment. In Garten's study, DR professionals related babies' distress to dyspnoea or agitation in this context.16 Given the specificity of the eol period in extremely preterm babies, tools validated for prolonged pain assessment in the NICU may not be fully appropriate.29
Surprisingly, administration of comfort medication was not influenced by pain assessment but only by the occurrence of gasping. Its use to stop gasping by dying neonates is not evidence-based and is hard to justify, as gasping respiration is known to be a brainstem autoresuscitation reflex secondary to asphyxia; it is characterised by markedly depressed brain function with a flat EEG30 ,31 and is not associated with the transmission of a painful stimulus. Neither opioids nor sedatives have been shown to reduce this sign.32 ,33 The decision to start medication in these cases of gasping or dyspnoea might therefore be related to the professional's own distress or to (supposed or real) parental distress.16 ,26 ,34 ,35 Eol symptoms have nonetheless been reported to be difficult for babies and distressing for medical staff.26 ,34 ,35 Parents may well be distressed after treatment withdrawal in the NICU by their baby's gasping respiration, which some may interpret as the baby saying: ‘I want to live’;36 a few cases have described parents' requests to stop their babies from gasping after extubation.17 ,18 In our study, NMB administration was reported in two cases, both with gasping, prolonged in one. This medication may have been intended to stop this symptom.
All but one of the babies in our study lived for <2 hours, a duration of life similar to that reported in unmedicated newborns.9 ,16 ,17 Although parents will be distressed by their child's protracted death,36 comfort medication administered to soothe the infant should not affect their duration of life, as it does not affect the time from treatment withdrawal to death in NICUs.17 The high frequency of gasping reported in our study (66% of neonates) and the variability in duration of life we found underline the importance of preparing parents and staff for the baby's possible respiratory symptoms (dyspnoea and gasping) and warning them about the uncertainty of duration of life.
The staff's satisfaction with care appeared to be associated with parent–baby contact, related perhaps to experiencing comfort care positively, as a ‘harmonious emotional parent–newborn interaction’.16 Contact was probably not routinely planned. In this context, a tactful proposal might give the parents the opportunity to decide themselves if they want to see or hold their baby37 as, at these early GAs, the mother does not always perceive the psychological individuality of the fetus as a child.38 Most mothers were offered psychological support and medical follow-up. These offers are helpful after perinatal loss because they imply recognition of loss to women at risk of disenfranchised grief and enable parents to access appropriate support later on, when their mourning feels socially unrecognised.10 Conversely, DR personnel were rarely offered psychological support, although their well-being is essential for them personally as individuals and also for the management the hospital can provide to its patients: burnout and learning to trivialise loss may affect the quality of the support they offer.39 ,40 Most were fairly satisfied with the care, probably because of the care and attention given to the baby and family. Less satisfaction with the care provided was reported by one third of the professionals and might be related to insufficient educational and/or emotional support, stress related to communication with parents16 or disagreement with the decision to withhold treatment, which could generate moral distress.41 Our results may serve as the starting point of a collaborative project aimed at improving the education and organisation of DR staff.
This study has limitations: data collection by questionnaires facilitates systematic and up-to-date gathering of information but can be associated with response bias, such as social desirability bias, whereby participants might tend to answer in a way that conforms to socially acceptable values.42 Anonymisation of the responses was sought to reduce this bias. Since all the participating centres volunteered to participate in this study, a selection bias cannot be ruled out, and generalisability to other centres should be cautious. Furthermore, the staff members' awareness of the questionnaire might have affected their clinical practice in these cases. Some missing data are another limitation; in particular, specifics about medication were insufficiently recorded; the baseline characteristics of the 10 infants without questionnaires were comparable with those of the infants analysed. Finally, the mechanisms by which local policies may have influenced the use of medications and of umbilical catheters are not known, and exploring the reasons for medication administration would require a qualitative study. Notwithstanding these limitations, the results of this multicentre study are important for they describe DR comfort care practices with a prospective design for the first time. These results point to pathways for future research: development of a specific distress tool adapted to the setting and patients' signs, and prospective assessment of the efficacy of both non-pharmacological treatment and minimally invasive routes for opioid administration. Finally, a qualitative exploration of parental and even perhaps staff experience with eol signs would be helpful, especially in developing guidelines for the use of comfort medication.
In conclusion, non-pharmacological comfort care was provided to nearly all babies dying in the DR; parents were also offered support. Half the babies received comfort medication, much more than in previous reports from DR settings. Medication was most often given to babies who were gasping. This underscores the need to interpret this sign properly in this context since gasping is not associated with suffering, whereas medication routes were frequently invasive. Future research should focus on standardised criteria for starting comfort medication for extremely premature infants dying in the DR, from the perspectives, studied as described above, of babies, their parents and healthcare providers.
Acknowledgments
The authors are grateful to the colleagues who contributed to this study. The authors are indebted to the parents who agreed to participate in the EPIPAGE-2 cohort study.
References
Footnotes
Collaborators Extreme Prematurity Writing Group: Elie Azria, Pierre Bétrémieux, Alain Beuchée, Pascal Boileau, Laurence Boujenah, Laurence Caeymaex, Frédérique Charlot, Olivier Claris, Anaëlle Coquelin, Caroline Diguisto, Micheline Garel, Laurence Foix L'Hélias, Monique Kaminski, Pierre Kuhn, Cerise Levaillant, Florence Rouget, Umberto Simeoni, Barthélemy Tosello, Patrick Truffert.
Contributors XD participated in the data analysis and interpretation and writing of the manuscript. CS-L and LB participated in the data collection, interpretation and writing of the manuscript. PB, MG, LFLH and MK initiated and designed the protocol, and participated in the data interpretation and writing of the manuscript. OC participated in writing the manuscript. LC initiated and designed the protocol, participated in the data collection and interpretation and in writing the manuscript. All the co-authors approved the final manuscript.
Funding This work was supported by a grant from the Fondation de France (http://www.fondationdefrance.org), which had no role in the design and conduct of the study, the collection, analysis or interpretation of data or writing of the report.
Competing interests None declared.
Ethics approval Consultative Committee on the Treatment of Information on Personal Health Data for Research Purposes (no 10.626); Committee for the Protection of People participating in Biomedical Research (no CPP SC-2873); The National Data Protection Authority (CNIL no 911009).
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The original version of the questionnaire is available to readers by asking the corresponding author.