Withdrawal and limitation of life-sustaining treatments in a paediatric intensive care unit and review of the literature

Objectives:   To examine withdrawal and limitation of life-sustaining treatment (WLST) in an Australian paediatric intensive care unit (PICU) and to compare this experience with published data from other countries.
Design:   Retrospective chart review and literature review.
Source of Data:   Review of 12 months of patient records from a tertiary Australian children's teaching hospital. Medline search using relevant key words focusing on death and PICU.
Results:   Twenty of 27 deaths (74%) followed either WLST ( n  = 16) or Do Not Resuscitate (DNR) orders ( n  = 4); five children failed cardiopulmonary resuscitation (CPR); and two children were brain-dead. Meetings between the medical team and family were documented for 15 of 16 children (93.8%) before treatment was withdrawn. The average time between withdrawal of life support and death was 13 min. A review of the English-language literature revealed that 18–65% occurring in PICUs worldwide follow WLST and/or institution of DNR orders. Rates were higher (30–65%) in North America and Europe than elsewhere. Most PICU deaths occurred within 3 days of admission. North American and British parents appear to be involved in decisions regarding withdrawal and limitation of treatment more often than parents in other countries.
Conclusions:   Withdrawal and limitation of life-sustaining treatment was more common in an Australian children's hospital ICU than has been reported from other countries. Details of discussion with parents, including the basis for any decision to WLST, were almost always documented in the patient's medical record.     

Variables influencing end-of-life care in children and adolescents with cancer

BACKGROUND: The purpose of this study was to describe the variables influencing end-of-life care in children and adolescents dying of cancer. MATERIALS AND METHODS: Records of 146 children with cancer who died at Children's Hospital were reviewed for demographics, diagnosis, location of death, withdrawal of life support, use of "do not resuscitate" (DNR) orders, and the length of time that those orders were in effect. RESULTS: Ninety-five patients were evaluated. Fifty-nine died of progressive disease and 36 deaths were therapy-related. Sixty-four percent of disease-related deaths occurred at home with support from home care or hospice. Only 10% of all patients died while receiving maximal aggressive support in the intensive care unit. Age, diagnosis (solid tumor vs. leukemia), cause of death, length of last hospital admission, and the duration of DNR orders had a significant correlation with the place of death and referral to and use of hospice. Thirty-five percent of all patients had hospice support. CONCLUSIONS: Most children who die of cancer die because of progressive disease at home with hospice support. Do not resuscitate orders were written for most patients who died. End-of-life decisions are influenced by patient diagnosis, cause of death, and age.     

Do Not Resuscitate orders and ethical decisions in a neonatal intensive care unit in a Muslim community

AIMS: To evaluate the need for Do Not Resuscitate (DNR) orders in a tertiary referral centre for neonatal intensive care, the criteria used in making these decisions, and the applicability of the Muslim ethical stance among parents in an Islamic community. METHODS: A prospective evaluation of all DNR decisions in the neonatal intensive care unit at the Royal Hospital in Oman, over a one year period between November 1999 and October 2000. This included decision criteria, and parental responses and expectations. RESULTS: Of 659 admissions to the neonatal intensive care unit during this period, DNR orders were written in 39 (6%) instances. Most related to congenital malformations (24/39, 62%). In those in whom ventilation was commenced (19/39, 49%) withdrawal was not culturally acceptable and expressly permitted in only 11%. For those in whom ventilation was not commenced (20/39, 51%), 70% agreed not to put their child on the ventilator if they did require it. Presence of extended family support (grandparents) and clergy was extremely useful. CONCLUSIONS: Asking parents alone to be explicitly involved or take full responsibility for decisions involving life and death is not culturally or socially acceptable in this community. Presence of extended family, and indirectly sounding out and taking into account their wishes, is more appropriate after assessing the resources and support services available.     

Circumstances of dying in hospitalized children

Conditions of dying in a tertiary children's hospital were assessed in a retrospective cohort study. Non-survivors, excluding newborns and emergency room patients, were allocated to four groups: brain death (BD), failed cardiopulmonary resuscitation (failed CPR), death following a do-not-resuscitate (DNR) order and death following withholding or withdrawal of therapy (W/W). In a 4-year period 190 (1.3%) of 14,903 admitted patients died. Of these 134 (71%) died on the paediatric intensive care unit, 42 (22%) on the ward and 14 (7%) in the operating room. W/W was found in 75 (39%), failed CPR in 57 (30%), BD in 32 (17%), and death following a DNR order in 26 (14%). Justifications for restrictions of treatment (W/W or DNR) were imminent death in 41 (41%), lack of future relational potential in 13 (13%) and excessive burden of disease in 47 (47%). In non-survivors analgesics and sedatives were frequently used to relieve suffering in the terminal phase. General principles for the approach of terminally ill children in whom death may become an option instead of a fate are discussed. Conclusion In the majority of children dying in hospital, death occurred following restrictions of life-sustaining treatment, comprising do-not-resuscitate or other forms of withholding or withdrawal of therapy. Received: 24 December 1997 / Accepted in revised form: 30 May 1998     

Measuring maternal stress and perceived support in 25 Italian NICUs

Aims: To determine the validity and reliability of the Parental Stressor Scale: Neonatal Intensive Care Unit (PSS: NICU) and the Nurse Parental Support Tool (NPST) for use with Italian parents; to investigate to which extent demographic variables and/or situational factors affect NICU‐related maternal stress. Methods: Mothers (N = 156) of very preterm (VPT) infants from 25 Italian NICUs completed a socio‐demographic form, the PSS: NICU and the NPST at discharge. Psychometric properties of both tools were evaluated. Results: High internal consistency and split‐half reliability were found for both measures. The multi‐dimensional structure of the PSS:NICU was confirmed. Alteration in parental role emerged as the greatest source of NICU‐related stress. Length of stay in NICU and familiar socio‐economic status explained partial variance in the PSS: NICU scores. NPST score mitigates the stress because of the infant’s appearance and behaviour, but not that related to the parental role alteration. Conclusions: PSS: NICU and NPST demonstrated adequate psychometric properties in a large sample of Italian mothers. The need for a psychologically informed support to NICU mothers is suggested.     

Physician end-of-life decision-making in newborns in a less developed health care setting: insight in considerations and implementation

BACKGROUND: A substantial proportion of the decisions to withhold or withdraw life-prolonging treatment are based on the newborn's predicted poor quality of life. All previous studies on end-of-life decisions were done in countries with adequate support for disabled neonatal intensive care units (NICU) survivors. Data on quality-of-life considerations in countries with developing health care are not available yet. AIM: The aim of the study was to examine the considerations of physicians taking end-of-life decisions in sick newborns and how those decisions are carried out in practice in a less developed health care setting. METHOD: Thirty-two deaths over 18 months in a neonatal unit were retrospectively analyzed. RESULTS: Twenty-four deaths (75%) were attributable to withholding or withdrawing of treatment. In 7 of these cases (29%), the decisions were based on quality-of-life considerations, mostly predicted suffering and expected hospital dependency. For the majority of paediatricians, end-of-life decision making was not influenced by legal or economic considerations or by considerations regarding availability of supportive care after discharge. CONCLUSION: Our study suggests that physician end-of-life decision making in this unit in a less developed health care setting is found to be similar to that in developed health care settings and is independent of availability of supportive care after discharge for infants with disabilities.     

Risk indicators for hearing loss in infants treated in different Neonatal Intensive Care Units

Aim: To assess which infants’ characteristics and specialized procedures are risk indicators for unilateral or bilateral hearing loss (HL) and to evaluate whether these risk indicators are associated with variation in prevalence of HL between Neonatal Intensive Care Units (NICUs). Methods: For 2002–2005, data from the NICU hearing screening database in the Netherlands were matched with the national neonatology database in which all NICU infants with their patient characteristics and specialized procedures are registered. Multivariate logistic regression analyses were performed to assess risk indicators for HL and to explain differences in prevalence rates between NICUs. Results: A total of 10 830 infants were available for analyses. The prevalence of HL was 1.8% and ranged from 0.7 to 3.7% between NICUs. Infants’ characteristics that significantly increased the risk of HL were the presence of craniofacial anomalies, chomosomal/syndromal anomalies, central nervous system conditions, circulatory system conditions and intra‐uterine infections. The specialized procedures involving ≥12 days of intensive care and high frequency oxygenation ventilation were independent risk indicators for HL. Approximately 20% of the variance can be explained by the studied risk indicators. Differences in prevalence rates between NICUs were slightly reduced after adjustment for these risk indicators. NICUs with the highest prevalence rates of HL were situated in the largest cities in the Netherlands with a mixed population because of immigration. Therefore, ethnicity may be a risk indicator. Conclusions: Several independent risk indicators for HL were found, but they could not explain all differences in prevalence rates of HL between NICUs.     

International study of wheezing in infants: risk factors in affluent and non‐affluent countries during the first year of life

Garcia‐Marcos L, Mallol J, Solé D, Brand PLP and EISL group. International study of wheezing in infants: risk factors in affluent and non‐affluent countries during the first year of life.
Pediatr Allergy Immunol 2010: 21: 878–888.
© 2010 John Wiley & Sons A/S Risk factors for wheezing during the first year of life (a major cause of respiratory morbidity worldwide) are poorly known in non‐affluent countries. We studied and compared risk factors in infants living in affluent and non‐affluent areas of the world. A population‐based study was carried out in random samples of infants from centres in Latin America (LA) and Europe (EU). Parents answered validated questionnaires referring to the first year of their infant’s life during routine health visits. Wheezing was stratified into occasional (1–2 episodes, OW) and recurrent (3 + episodes, RW). Among the 28687 infants included, the most important independent risk factors for OW and RW (both in LA and in EU) were having a cold during the first 3 months of life [OR for RW 3.12 (2.60–3.78) and 3.15 (2.51–3.97); population attributable fraction (PAF) 25.0% and 23.7%]; and attending nursery school [OR for RW 2.50 (2.04–3.08) and 3.09 (2.04–4.67); PAF 7.4% and 20.3%]. Other risk factors were as follows: male gender, smoking during pregnancy, family history of asthma/rhinitis, and infant eczema. Breast feeding for >3 months protected from RW [OR 0.8 (0.71–0.89) in LA and 0.77 (0.63–0.93) in EU]. University studies of mother protected only in LA [OR for OW 0.85 (0.76–0.95) and for RW 0.80 (0.70–0.90)]. Although most risk factors for wheezing are common in LA and EU; their public health impact may be quite different. Avoiding nursery schools and smoking in pregnancy, breastfeeding babies >3 months, and improving mother’s education would have a substantial impact in lowering its prevalence worldwide.     

Parents, siblings and grandparents in the Neonatal Intensive Care Unit A survey of policies in eight European countries

Objective:  To describe policies towards family visiting in Neonatal Intensive Care Units (NICU) and compare findings with those of a survey carried out 10 years earlier.
Methods:  A questionnaire on early developmental care practices was mailed to 362 units in eight European countries (Sweden, Denmark, the UK, the Netherlands, Belgium, France, Spain and Italy). Of them 78% responded, and among those responded, 175 reported caring for at least 50 very low birth weight infants every year and their responses were analysed further.
Results:  A majority of all units allowed access at any time for both parents. This was almost universal in northern Europe and the UK, whereas it was the policy of less than one-third of NICU in Spain and Italy, with France in an intermediate position. Restrictions on visiting of grandparents, siblings and friends, as well as restricting parents' presence during medical rounds and procedures followed the same pattern. A composite visiting score was computed using all the variables related to family visiting. Lower median values and larger variability were obtained for the southern countries, indicating more restrictive attitudes and lack of national policy.
Conclusions:  The presence of parents and other family members in European NICUs has improved over a 10-year period. Several barriers, however, are still in place, particularly in the South European countries.     

Decision making and modes of death in a tertiary neonatal unit

AIMS: To study the frequency and reason for withdrawal/withholding of life sustaining treatment (LST) and do not resuscitate (DNR) orders in infants who died in a tertiary neonatal unit. METHODS: Infants who died at Homerton University Hospital between January 1998 and September 2001 were studied by retrospective analysis of patient records. RESULTS: The case notes of 71 (84%) of 85 infants who died were studied. Mode of death was withdrawal of LST in 28 (40%), DNR in 11 (15%), withholding of LST in two (3%), and natural in 30 (42%) infants. Withdrawal of LST was discussed with the parents of 39 seriously ill infants; 28 (72%) parents agreed. There was no difference in birth weight and gestational age of babies whose parents agreed or refused withdrawal of LST. White and Afro-Caribbean parents and those from the Indian subcontinent (20 of 23) were more likely to agree to withdrawal of LST than Black African or Jewish (eight of 16, p = 0.015) parents. The median age at withdrawal of LST was 4 days (range 1-57). The median duration between discussion and the parents agreeing to withdrawal of LST was 165 minutes (range 30-2160), and median duration between withdrawal of LST and death was 22 minutes (range 5-210). The most common reason for withdrawal of LST was complications of extreme prematurity (68%). CONCLUSION: The most common mode of death was withdrawal of LST, and the most common reason was complications of extreme prematurity. The ethnic and cultural background of the parents influenced agreement to withdrawal of LST.     

Exploring Acceptance and Commitment Therapy for parents of preterm infants

The start of a parenting journey in the neonatal intensive care unit (NICU) presents many stressors to parents. Previous research has shown parents of infants admitted to the NICU experience heightened stress, anxiety, and depression. Mental health support varies across Canadian NICUs with mixed results. One promising intervention that has not been explored in the NICU is Acceptance and Commitment Therapy (ACT), a behavioural therapy that has had positive mental health-related outcomes in similar parental populations. ACT differs from previous mental health interventions such as traditional Cognitive Behavioural Therapy (CBT) as it involves mindfulness and acceptance to increase psychological flexibility. Increased psychological flexibility is linked to greater emotional well-being, a higher quality of life, and decreased stress, anxiety, and depression. There is a need for research investigating the utility of ACT in improving mental health outcomes for parents of preterm infants.     

Death in the neonatal intensive care unit: changing patterns of end of life care over two decades

BACKGROUND: Death remains a common event in the neonatal intensive care unit, and often involves limitation or withdrawal of life sustaining treatment. OBJECTIVE: To document changes in the causes of death and its management over the last two decades. METHODS: An audit of infants dying in the neonatal intensive care unit was performed during two epochs (1985-1987 and 1999-2001). The principal diagnoses of infants who died were recorded, as well as their apparent prognoses, and any decisions to limit or withdraw medical treatment. RESULTS: In epoch 1, 132 infants died out of 1362 admissions (9.7%), and in epoch 2 there were 111 deaths out of 1776 admissions (6.2%; p<0.001). Approximately three quarters of infants died after withdrawal of life sustaining treatment in both epochs. There was a significant reduction in the proportion of deaths from chromosomal abnormalities, and from neural tube defects in epoch 2. CONCLUSIONS: There have been substantial changes in the illnesses leading to death in the neonatal intensive care unit. These may reflect the combined effects of prenatal diagnosis and changing community and medical attitudes.     

“DO NOT RESUSCITATE” ORDERS AMONG CHILDREN WITH SOLID TUMORS AT THE END OF LIFE

The goal of this study was to evaluate the frequency and timing of “do not resuscitate” (DNR) orders among pediatric oncology patients with progressive solid tumors. A retrospective evaluation of the medical charts of 36 patients who died in our department over the last 4 years was made. There were 21 males and 15 females with a mean age of 10 years (range, 1–22 years). Fourteen patients had brain tumors, 12 had sarcomas, 5 had lymphomas, 4 had neuroblastomas, and 1 suffered from hepatocellular carcinoma. A DNR order was registered in the medical charts of 22 (61%) patients. The DNR was ordered within less than 24 h of death in 5 patients. Median time from DNR order until death was 5 days (range, 0.25–82 days). Place of death was home for 5 (14%) patients, 3 of whom had DNR orders; pediatric oncology ward for 28 (78%), 19 of whom had DNR orders); ICU for 3 (8%), none with a DNR order. Mean time from last day of anticancer treatment until death was 63 days in the group with DNR orders and 56.5 days in the group without DNR orders (p = NS). The study showed that (1) DNR orders were written in only about half the charts of patients with progressive cancer; (2) in several cases, DNR orders were given close to death; and (3) advanced discussion of DNR with parents is needed to reduce parental and medical staff stress accompanying the death of a child and optimizing management of the terminal phase of the disease.     

Delivery room decision-making at the threshold of viability

OBJECTIVES: To assess attitudes of neonatologists toward parental wishes in delivery room resuscitation decisions at the threshold of viability. STUDY DESIGN: Cross-sectional survey of the 175 practicing level II/III neonatologists in six New England states. RESULTS: Response rate was 85% (149/175). At 24 1/7-6/7 weeks' gestation, 41% of neonatologists considered treatment clearly beneficial, and at 25 1/7-6/7 weeks' gestation, 84% considered treatment clearly beneficial. When respondents consider treatment clearly beneficial, 91% reported that they would resuscitate in the delivery room despite parental requests to withhold. At or below 23 0/7 weeks' gestation, 93% of neonatologists considered treatment futile. Thirty-three percent reported that they would provide what they consider futile treatment at parental request. When respondents consider treatment to be of uncertain benefit, all reported that they would resuscitate when parents request it, 98% reported that they would resuscitate when parents are unsure, and 76% reported that they would follow parental requests to withhold. CONCLUSIONS: Variation in neonatologists' beliefs about the gestational age bounds of clearly beneficial treatment and attitudes toward parental wishes in the context of uncertainty is likely to impact the manner in which they discuss options with parents before delivery. This supports the importance of transparency in neonatal decision-making.     

End-of-life decisions in delivery room and neonatal intensive care unit

Background: The increase in neonatal survival in recent decades has been followed by an increase in later disabilities. This has given rise to many new ethical issues. In different countries, efforts are being made to define ethical guidelines regarding withholding or withdrawing intensive care and end-of-life decisions in critically ill newborn infants. These guidelines have to be differentiated from ethical decision-making models which structure the process of decision making for an individual child. Such a framework has been in existence in our clinic for 10 years. Aim: The aims of this study were to evaluate how end-of-life decisions are taken in our perinatal centre and to analyse whether these decisions are consistent with our framework for structured ethical decision making. Methods: 199 consecutive neonatal deaths over 5 y were evaluated. Results: In 157 cases (79%), end-of-life decisions were taken according to our ethical framework; in the remaining 42 cases (21%), the baby died before this could be done. In 92% of cases, parents were involved in the decision and, in all cases but one, agreed with the decision. A patient's life was never intentionally and actively terminated.

Conclusion: In contrast to earlier years, in-hospital death in our clinic is nowadays usually preceded by structured and documented medical end-of-life decisions.     

Neonatal end-of-life practice in a German perinatal centre

AIM: To investigate the end-of-life practice in a large perinatal centre in Germany. METHODS: Retrospective chart review was performed in all neonates deceased in the delivery room (n = 31) and the neonatal intensive care unit (n = 47) between 2002 and 2004. RESULTS: Neonatal death was preceded by an end-of-life decision (EOLD) in 81% of cases in the delivery room and 83% in the neonatal intensive care unit. The majority of deceased neonates were born prematurely or with congenital malformation. Life-sustaining treatment was not initiated in 74% of the infants deceased in the delivery room. In the unit, 52% died after withdrawal of therapy. Mechanical ventilation was withdrawn most frequently (79% of cases). Futility and immediate death were common considerations in EOLD, but the infant's suffering and future quality of life also played a role. Parents were involved in EOLD-making in all but emergency cases. No active termination of life was performed. CONCLUSION: In our perinatal centre, the majority of neonatal deaths occurred after limitation of therapy. Treatment was actively withdrawn in half of the infants in the neonatal intensive care unit. Actual end-of-life practice in a large perinatal centre differs from the restrictive attitude towards EOLD reported for German neonatologists in previous surveys.     

End-of-life decision making in neonates and infants: comparison of The Netherlands and Belgium (Flanders)

AIM: We compared the results of two recent studies on end-of-life decisions (ELDs) for neonates and infants in Belgium (Flanders) and The Netherlands. METHODS: Questionnaires were sent to physicians who reported the death of a child under the age of 1 (Belgium: n = 292, response 87%; Netherlands: n = 249, response 84%). The questionnaires included structured questions about whether death had been preceded by ELDs, and about the decision-making process. RESULTS: In both countries, in about 25% of all deaths a life-sustaining treatment was withheld, and in about 40% pain or other symptoms were alleviated taking into account that death might be hastened. In Belgium, a life-sustaining treatment was less often withdrawn than in The Netherlands (32% vs. 50%, respectively). Drugs were administered with the explicit intention of hastening death in similar percentages of all deaths (Belgium: 7%; Netherlands: 9%). Dutch physicians more often than Belgian physicians discussed ELDs with parents (96% vs. 81%, respectively), and with colleague physicians (94% vs. 80%, respectively). CONCLUSIONS: End-of-life decision making in severely ill neonates seems to be rather similar in Belgium and The Netherlands. Differences are that Dutch physicians more often withdraw life-sustaining treatment. Furthermore, parents and colleague physicians are more often involved in the decision making in The Netherlands.     

Risk adjusted and population based studies of the outcome for high risk infants in Scotland and Australia. International Neonatal Network, Scottish Neonatal Consultants, Nurses Collaborative Study Group

OBJECTIVES—To compare outcomes of care in selected neonatal intensive care units (NICUs) for very low birthweight (VLBW) or preterm infants in Scotland and Australia (study 1) and perinatal care for all VLBW infants in both countries (study 2).
DESIGN—Study 1: risk adjusted cohort study; study 2: population based cohort study.
SUBJECTS—Study 1: all 2621 infants of < 1500 g birth weight or < 31 weeks'' gestation admitted to a volunteer sample of hospitals comprising eight of all 17 Scottish NICUs and six of all 12 tertiary NICUs in New South Wales and Queensland in 1993-1994; study 2: all 5986infants of 500-1499 g birth weight registered as live born in Scotland and Australia in 1993-1994.
MAIN OUTCOMES—Study 1: (a) hospital death; (b) death or cerebral damage, each adjusted for gestation and CRIB (clinical risk index for babies); study 2: neonatal (28 day) mortality.
RESULTS—Study 1. Data were obtained for 1628 admissions in six Australian NICUs, 775 in five Scottish tertiary NICUs, and 148 in three Scottish non-tertiary NICUs. Crude hospital death rates were 13%, 22%, and 22% respectively. Risk adjusted hospital mortality was about 50% higher in Scottish than in Australian NICUs (adjusted mortality ratio 1.46, 95% confidence interval (CI) 1.29 to 1.63,p < 0.001). There was no difference in risk adjusted outcomes between Scottish tertiary and non-tertiary NICUs. After risk adjustment, death or cerebral damage was more common in Scottish than Australian NICUs (odds ratio 1.9, 95% CI 1.5 to 2.5). Both these risk adjusted adverse outcomes remained more common in Scottish than Australian NICUs after excluding all infants < 28 weeks'' gestation from the comparison. Study 2. Population based neonatal mortality in infants of 500-1499 g was higher in Scotland (20.3%) than Australia (16.6%) (relative risk 1.22, 95% CI 1.08 to 1.39, p = 0.002). In a post hoc analysis, neonatal mortality was also higher in England and Wales than in Australia.
CONCLUSIONS—Study 1: outcome was better in the Australian NICUs. Study 2: perinatal outcome was better in Australia. Both results may be consistent, at least in part, with differences in the organisation and implementation of neonatal care.

     

Gliosis in neonatal SUDI cases

Aim: To review sudden unexpected infant deaths (SUDI) in the first 28 days of life referred to a Coronial Perinatal Forensic Pathology Service over a 10‐year period from 2000 to 2009. Methods: Cases were collected from mortuary records, and a retrospective review of autopsy reports and other available infant records was undertaken. Results: Twenty‐four neonatal SUDI were reviewed. For eight infants, a diagnosis was made at autopsy. For the remaining 16 infants, 14 (87.5%) were bedsharing at the time of death. Maori infants and those living in deprived neighbourhoods were over‐represented. Only two infants were preterm, and four were growth‐restricted. At post‐mortem, white matter gliosis was found in 10 of the 16 (62.5%) unexplained SUDI cases. Conclusion: Sudden unexpected infant deaths occur in the first month of life in association with bedsharing. Gliosis may be an important associated risk factor, and its presence indicates a previous insult of prenatal onset.     

Risk factors associated with pneumothorax in Malaysian neonatal intensive care units

Aims: This study aimed to determine the risk factors associated with the development of pneumothorax among infants admitted to the Malaysian neonatal intensive care units (NICUs). Methods: Twenty‐nine of forty NICUs in Malaysian public hospitals participated in the 2006 Malaysian National Neonatal Registry (MNNR). Data of infants from this registry with and without pneumothorax were analysed. A diagnosis of pneumothorax was made in the presence of extra‐pleural air detected by chest radiograph or needle aspiration. Results: There were 10 387 infants admitted to these NICUs who met the MNNR inclusion criteria and were included in this study. Pneumothorax developed in 505 (4.9%) of them. Pneumothorax was most common (7.3%) among the extremely low birthweight infants (birthweight equal or less than 1000 g) and the extremely preterm infants of gestation equal or less than 26 weeks (6.8%). Logistic regression analysis showed that the significant risk factors associated with increased risk of development of pneumothorax were: meconium aspiration syndrome (adjusted odds ratio (OR) = 2.1, 95% confidence intervals (CI): 1.7, 2.7), intermittent mandatory ventilation (adjusted OR = 1.5, 95% CI: 1.2, 2.0), high‐frequency oscillatory ventilation (adjusted OR = 3.9, 95% CI: 3.0, 5.2) and confirmed sepsis (adjusted OR = 1.6, 95% CI: 1.3, 2.1). Infants on nasal continuous positive airway pressure (nCPAP) therapy were associated with significantly lower risk of pneumothorax (adjusted OR = 0.5, 95% CI: 0.4, 0.6). Conclusion: Meconium aspiration syndrome, mechanical ventilation and sepsis were associated with increased risk and the use of nCPAP with decreased risk of pneumothorax in Malaysian NICUs.