Our aim was to evaluate programs promoting bed sharing on maternity wards and determining ways to reduce these risks.
Members of the National Association of Medical Examiners were contacted requesting information on deaths of healthy infants while bed sharing on maternity wards.
Fifteen deaths and three near deaths are reported. One or more factors that increase the risk of bed sharing were present in all cases. Accidental suffocation was deemed the most likely cause of these incidents.
Cases of infant deaths and near deaths while bed sharing on maternity wards are under reported. The ‘Baby Friendly’ (BF) initiative in maternity hospitals to promote breastfeeding is endorsed by the American Academy of Pediatrics and the US Center for Disease Control and Prevention. The BF initiative encourages prolonged skin-to-skin contact and bed sharing. Education of mothers and more efficient monitoring should significantly reduce the risk of maternity ward bed sharing.
Although bed sharing with infants is well known to be hazardous, deaths and near deaths of newborn infants while bed sharing in hospitals in the United States have received little attention aside from a recent brief report of a single death and two near deaths.1 In the recent past, similar deaths and near deaths have been reported in Europe and Britain.2, 3, 4, 5, 6, 7, 8, 9, 10, 11 The incidence of these events is reported to be as high as 4/1000 live births.12 These events occurred within the first 24 h of birth during ‘skin-to-skin’ contact between mother and infant, a practice promoted by the ‘Baby Friendly’ (BF) initiative endorsed by United Nations Children’s Fund, BF, USA and the US Center for Disease Control and Prevention to support breastfeeding.13, 14, 15, 16, 17 We report 15 deaths and 3 near deaths of healthy infants occurring during skin-to-skin contact or while bed sharing on maternity wards in the United States. Our findings suggest that such incidents are underreported in the United States and are preventable. We suggest ways in which close maternal infant contact to promote breast feeding may be done more safely.
In 2011, information on deaths of healthy newborn infants while bed sharing in hospital was requested by email from members of the National Association of Medical Examiners using a listserv. Persons on the Center for Disease Control and Prevention Sudden Unexpected Infant Death listserv were also contacted. Information with identifiers deleted on 15 deaths was forwarded. Information on three near deaths of infants, although not originally requested, was forwarded to us by close colleagues and others who were aware of our interest in such cases. The information we obtained included post mortem reports, hospital records and other sources (Table 1). This research was approved by an institutional committee for human research.
Data for the cases are shown in Table 2. All infants were presumed to be healthy at the time the incident occurred. The incidents occurred between 1999 and 2013. The infants mean age was 23.9 h (range=1.15 h to 3 days). The median age was 14.75 h (±5 and 45.75 h). The mother’s pregnancy and delivery were minimally complicated in 10 cases (Table 2). Of the 13 cases where mother’s parity was known, 7 were primipara and 6 were multipara. Apgar scores available in 16 cases were all within normal limits. Medical examiners’ or physicians’ diagnoses included overlaying (one case), sudden infant death syndrome (SIDS) (one case), cause of death undetermined (eight cases), accidental suffocation (five cases) and apnea of undetermined etiology (three cases). Two near deaths resulted in severe lasting brain injury.
In eight cases, the mother fell asleep while breastfeeding (Table 2). In four cases, the mother woke up from sleep but believed her infant to be sleeping when an attendant found the infant lifeless. One or more risk factors that are known or suspected (obesity and swaddling) to further increase the risk of bed sharing were present in all cases.17, 18, 19, 20, 21 These included the infants’ age <4 months in all 17 cases; maternal sedating drugs in 7 cases; cases excessive of maternal fatigue, either stated or assumed if the event occurred within 24 h of birth in 12 cases; pillows and/or other soft bedding present in 9 cases; obesity in 2 cases; maternal smoking in 2 cases; and infant swaddled in 4 cases (numbers 6, 10, 12 and 16).
Factors presumed to potentially reduce the risk of bed sharing were present in most of the cases. These included mother awake when the death or near death occurred in 4 cases and other individuals in the room when the event occurred in 10 cases.
This study suggests that the majority or all of the deaths and near deaths reported here were preventable. In six infants, routine screens for metabolic inborn errors were reported as normal (Table 2). For these and the other infants, evidence from the scene indicated suffocation as the most likely cause of death. Therefore, it is unlikely that inherited metabolic disorders had a role in the cases reported here.
In spite of the reports from overseas, it is likely that most maternity hospital personnel and neonatologists are still unaware of the risk of suffocation during close maternal infant contact. In the three near-death cases accidental suffocation was not considered in the differential diagnosis, and in two cases (numbers 17 and 18) the mother was initially accused on totally insufficient grounds of intentionally suffocating her infant. In one of these cases (number 17), the incident was allegedly reported to a child protection agency.
In the European literature, several authors suggest the causal mechanism resulting in the fatal or near-fatal incidents is airway occlusion when the infant’s face is in contact with the mother’s breast or abdomen.2, 9 In our study, eight infants were ‘on the breast’ when discovered. Occlusion of an infant’s airway while breastfeeding has been previously reported as a cause of death.22, 23 Relief from obstruction depends on the mother’s responses and/or the infant’s neck extension reflex initiated by airway occlusion.24 Breastfeeding in a recumbent position may increase the risk of airway obstruction if the infant’s attempt to withdraw from the breast is blocked by the mother’s hand or arm.
The incidents reported here are probably a substantial underestimate of such occurrences in past years. We did not request information on near deaths. The three cases reported here were brought to our attention by individuals aware of our interest in such cases. Previous reports of near deaths were nearly twice that of reports of actual deaths.9 Furthermore, in four instances information on deaths were not reported to us by National Association of Medical Examiners members but were brought to our attention by parents or others. In four additional deaths, information was requested but this was not made available to us. In order to know the true incidence of such deaths and near deaths, we feel that in the future reporting to a public health agency should be mandatory as is already the case for other adverse in-hospital events.
Whereas previous reports have been concerned with adverse incidents during the first 24 h of life, the present study indicates that infants are at risk for death or injury throughout the hospital stay. A stated aim of BF USA is to ‘help mothers initiate breastfeeding within one hour of birth’.14 To facilitate this, the American Academy of Pediatrics recommends direct skin-to-skin contact with mothers immediately after delivery and as much as possible throughout the post-partum period.15, 25 BF USA advises that infants and mothers share a room continuously and that infants be breast fed on demand without restricting the duration of the feeding and with a minimum of 10–12 feedings in 24 h.14 In addition United Nations Children’s Fund encourages in-hospital bed sharing.26 These recommendations will likely result in bed sharing for prolonged periods of time, particularly for mothers who have had cesarean sections or others, who do not wish to leave their beds frequently to breastfed.
Some professionals see minimal risks in bed sharing, with few exceptions, and this may explain a nurse not taking time to remove the infant from the mother’s bed after being requested to do so (case number10).27, 28 One hundred fifty hospitals in the United States are already certified ‘BF,’ and reportedly many more are seeking to become certified (http://www.babyfriendlyusa.org/find-facilities). The American Academy of Pediatrics and the Center for Disease Control and Prevention have taken steps to increase the number of BF hospitals in the United States.14, 15, 16, 17 This stems from the findings that skin-to-skin contact has been shown to be beneficial for mother–infant bonding, and the BF approach reportedly increases breastfeeding rates after hospital discharge.29, 30
An important question is: can in-hospital bed sharing be done safely? The presence of other individuals in the room could reduce the risk, but in 10 instances resulting in death or near death other individuals were present. In three of the present cases, the room was dark or dimly lit. This was likely in many of the other cases, a situation hindering the ability to assess the infant’s wellbeing. In addition, the finding that in four cases the mother was awake at the time of the incident indicates that wakefulness is not completely protective. It has been previously reported that in 20% of death and near deaths, the mother was awake when the incident occurred.10 In the United States, breastfeeding during bed sharing is regarded as safe so long as the mother is awake and the United Nations Children’s Fund UK BF initiatives concurs with this.18, 30 The present and prior reports clearly show this is not the case during the immediate post-partum period. Frequent bed checks by nurses at 5–10 min intervals is advocated in the BF literature, but as deaths or injury can occur in a few minutes, such checks would have to become nearly continuous to entirely prevent death or injury.13
We have several suggestions to increase the safety of infants on maternity wards. First, hospital personnel and expectant parents should be made aware of potential hazards of skin-to-skin and bed-sharing practices. Mothers should be taught to access breathing, skin color and response to stimuli in their infants, and when in close contact there should be sufficient light in the room to allow assessment. When a mother is in close contact with her infant, one-on-one supervision of infant and mother should be undertaken by a person trained to monitor the infant’s wellbeing as well as the mother’s wakefulness. In many cases, nurses will be unavailable for these duties. A previous recommendation is that close surveillance is needed especially during the first 2 h after delivery; however, the present findings indicate that close supervision is needed throughout the hospital stay when infants are in close contact with their mothers. In some cases, dedicated relatives or friends might perform this function. An alternative approach would be to electronically monitor infants (heart rate or arterial saturation) with alarms referred to the nursing station to avoid disturbing parents with false alarms. This would offer considerable protection for infants in close physical contact with mothers.
In summary, the BF initiative is an important public health program benefiting infants in many ways, including reducing the incidence of childhood obesity.16, 17 However, any extensive health program needs to be monitored for adverse effects and modified accordingly.
This study was funded by The Madison Foundation. Clifford C. Nelson, MD, for helpful suggestions as well as the medical examiners and infants' families for providing information, and Kathleen Harris for expert assistance in manuscript preparation.
About this article
Journal of Perinatology (2014)