Introduction
This retrospective study evaluated all consecutive pediatric and neonatal deaths recorded over a 48-month period between January 2008 and December 2010 at Sainte-Justine Hospital, a tertiary university-affiliated maternal and child healthcare center. The hospital includes a specialized delivery unit managing nearly 4,000 live births annually, with approximately half involving high-risk pregnancies. The institution also contains two intensive care units: a 70-bed neonatal intensive care unit (NICU) admitting roughly 1,400 newborns each year and a 24-bed pediatric intensive care unit (PICU) receiving nearly 1,000 admissions annually. In addition, an academic pediatric palliative care program actively supports around 100 patients and coordinates end-of-life care for approximately 50 terminally ill children every year.
The study also incorporated children who died at home or in hospice care while being closely monitored by the hospital’s pediatric palliative care team. Several exclusion criteria were applied, including stillbirths, pregnancy terminations, births before 23 weeks of gestation, deaths occurring during transport to the hospital, and patients who died in another institution despite previous follow-up at Sainte-Justine Hospital. Neonates who died before one month of age without being discharged after birth were categorized as neonatal deaths, even if death occurred outside the ICU setting.
Medical records were systematically reviewed to collect demographic and clinical information. For ICU-related deaths, investigators specifically examined the week preceding death to analyze end-of-life decisions and therapeutic limitations. Deaths were categorized according to physiological condition and medical intervention strategies, following previously established ethical classification models.

Classification of End-of-Life Decisions
The investigators divided deaths into four principal ethical categories:
- Death after ICU admission was withheld.
- Death despite full cardiopulmonary resuscitation (CPR) and continued life-sustaining intervention (LSI).
- Death during mechanical ventilation without CPR.
- Death following withdrawal or withholding of life-sustaining interventions.
Patients in the fourth category were further divided into physiologically stable or unstable groups before the final end-of-life decision. Instability criteria included severe respiratory failure despite maximal oxygen therapy, persistent hypotension requiring multiple vasopressors, prolonged bradycardia, refractory metabolic acidosis with pH ≤7.0, multiple organ dysfunction, or confirmed brain death.
Children who did not meet these severe physiological criteria were classified as stable, regardless of whether they remained mechanically ventilated. Two independent investigators reviewed and validated all classifications to ensure consistency.
Assessment of ICU Admission Decisions
For children who died after ICU admission was withheld, additional prognostic indicators were analyzed, including:
- predicted one-year survival probability,
- estimated long-term neurological outcome,
- potential requirement for lifelong mechanical ventilation,
- presence of severe encephalopathy,
- probability of intact five-year survival.
For premature infants, survival and morbidity predictions were estimated using the Eunice Kennedy Shriver National Institute of Child Health and Human Development preterm outcome calculator.
Statistical analyses were conducted using SPSS version 16. Comparative analyses between neonatal and pediatric populations were performed using descriptive statistics and chi-square testing, with statistical significance defined as P < 0.05.
Results of the Study
Overall Mortality
During the study period, 220 deaths were documented, including 99 neonatal deaths. Among these:
- 145 deaths occurred in intensive care units,
- 77 occurred in the NICU,
- 68 occurred in the PICU.
An additional 75 children died outside the ICU environment, including deaths occurring in delivery rooms, emergency departments, pediatric wards, hospice facilities, and home-care settings.
Only 6% of all deaths were preceded by active cardiopulmonary resuscitation, demonstrating that aggressive resuscitation was relatively uncommon before death in both neonatal and pediatric critical care settings.
Characteristics of Neonatal Deaths
NICU deaths occurred at a mean gestational age of 31.8 weeks, with nearly half involving extremely premature infants born before 29 weeks of gestation. The average birth weight was approximately 1,725 grams.
Major causes of neonatal mortality included:
- congenital malformations,
- necrotizing enterocolitis,
- neonatal sepsis,
- birth asphyxia,
- respiratory failure,
- intraventricular hemorrhage,
- complications related to prematurity.
The median age at death among neonates was eight days.
Characteristics of Pediatric ICU Deaths
Children who died in the PICU had an average age of 4.8 years, with many already suffering from severe chronic diseases before ICU admission. Common diagnoses included:
- congenital anomalies,
- cardiac diseases,
- severe trauma,
- sepsis and infections,
- neoplastic disorders,
- intracranial hemorrhage,
- brain death after traumatic injury.
The average PICU stay before death was approximately 13 days.
Differences Between NICU and PICU End-of-Life Practices
One of the most important findings of the study was the major ethical and clinical difference between neonatal and pediatric intensive care approaches.
PICU Death Patterns
In the PICU, most deaths occurred in physiologically unstable children receiving maximal intensive care. Mechanical ventilation was frequently maintained until death, and many patients died despite aggressive treatment.
These children often had:
- chronic life-limiting diseases
- repeated ICU admissions
- prolonged critical illness
- progressive organ failure
NICU Death Patterns
In contrast, the NICU showed a significantly higher proportion of deaths among physiologically stable infants after withdrawal or withholding of life-sustaining treatment due to concerns regarding long-term quality of life.
Many neonates who died in the NICU could potentially have survived with continued intensive intervention. However, decisions were frequently influenced by:
- predicted severe disability,
- neurological prognosis,
- anticipated poor quality of life,
- parental preferences,
- ethical considerations surrounding extreme prematurity.
Mechanical ventilation was commonly removed before death, allowing infants to die in their parents’ arms.
Ethical Differences in Neonatal and Pediatric Decision-Making
The study highlights a major ethical distinction between neonatal and pediatric medicine.
For neonates, especially extremely premature infants, healthcare teams and parents appeared more willing to:
- withhold ICU admission,
- limit intensive interventions,
- prioritize comfort care,
- base decisions on future disability risk.
In contrast, older children with established disabilities or chronic illnesses were more likely to continue receiving aggressive intensive care, even when prognosis remained poor.
This discrepancy suggests that ethical perceptions regarding disability, personhood, prognosis, and quality of life differ substantially between newborns and older pediatric patients.
Potential Explanations for Ethical Differences
Several factors may explain these contrasting approaches:
1. Prognostic Uncertainty
Older children may already have survived previous critical illnesses, making poor prognostic predictions less certain.
2. Emotional and Social Attachment
Parents and clinicians may develop stronger interpersonal bonds with older children, influencing decisions toward continued treatment.
3. Nature of Neonatal Disease
Many neonatal complications are directly associated with prolonged intensive care interventions, leading clinicians to perceive themselves as partially responsible for future disabilities.
4. Historical and Cultural Factors
Historically, neonatal mortality was relatively common, potentially shaping societal perceptions that infant death is more “natural” compared with the death of an older child.
Documentation and End-of-Life Communication
The study demonstrated that physicians consistently documented end-of-life discussions and ethical reasoning in medical records. Neonatologists frequently used terminology such as:
- “poor prognosis,”
- “quality of life,”
- “comfort care.”
Pediatric intensivists more commonly emphasized:
- chronic illness trajectory,
- treatment burden,
- medical instability.
This reflects important cultural differences between neonatal and pediatric critical care environments.
Clinical and Ethical Implications
Modern pediatric intensive care has dramatically improved survival rates for critically ill children who would previously not have survived. As a result, ethical decisions regarding when to initiate, continue, withhold, or withdraw life-sustaining therapy have become increasingly complex.
The study emphasizes that:
- ethical neutrality is difficult to achieve,
- quality-of-life judgments influence decision-making,
- disability risk strongly impacts neonatal care choices,
- ICU culture shapes end-of-life practices.
Understanding these underlying ethical frameworks is essential for improving consistency, transparency, family communication, and compassionate pediatric care.
Conclusion
This study demonstrates significant differences between neonatal and pediatric end-of-life decision-making practices. In neonatal intensive care, withdrawal or withholding of life-sustaining treatment often occurs in physiologically stable infants due to anticipated disability or poor quality of life. Conversely, pediatric intensive care deaths more frequently occur after prolonged aggressive treatment in unstable children with severe chronic disease.
These findings reveal the complex interaction between medical prognosis, ethical reasoning, cultural perceptions of disability, and emotional factors in pediatric end-of-life care. Continued discussion regarding ethical consistency and family-centered decision-making remains essential for improving care for critically ill children and neonates.


