Less Than Meets the Eye: The Consumer Product Safety Commission's Campaign Against Bed‐Sharing with Babies

In a surprisingly assertive and opinionated article last year, the United States Consumer Product Safety Commission (CPSC) took the field against the common practice of adults sharing beds with infants and children ( 1). At the press conference announcing their study, Ann Brown, the commission's chairwoman, was emphatic: “Don't sleep with your baby or put the baby down to sleep in an adult bed” ( 2). They recognized that their advice would be controversial, “But it would have been criminal of us to have this information and not bring it out” ( 3). The authors of a related study using much of the same data had the same message: “Bed-sharing and use of adult beds for infants should be discouraged” ( 4). The commission's expectation that the studies would provoke controversy was correct. Proponents of the “family bed”, experts on sudden infant death syndrome, professionals interested in sleep patterns, and advocates of breastfeeding all raised objections, and untold numbers of parents who sleep with their babies became anxious and agitated ( 2,3,5–11). The major subject of both studies was the risk to infants of sleeping in adult beds, with particular attention to the dangers of bed-sharing by adults and babies. The authors attempted to quantify the risk of death due to entrapment between mattress and bed frame or wall, entrapment in bed railings, suffocation in a soft bed or pillow, and smothering by overlying by an adult or an older child. Using three CPSC databases, Nakamura et al found 515 deaths of babies under 2 years of age in adult beds from 1990 to 1997, of which 394 were thought to have been caused by entrapment and 121 by overlying by an adult or a sibling. Cases diagnosed as sudden infant death syndrome that had occurred in adult beds or associated with bed-sharing were excluded. Using the basic CPSC Death Certificate File (DTHS), Drago and Dannenberg reported 2178 suffocation deaths of infants under 13 months from 1980 to 1997, of which 1263 were due to suffocation in a bed or crib, and 169 were ascribed to overlying. They also omitted any data on deaths in adult beds thought to have been due to sudden infant death syndrome. The first thing the commission should have admitted was that their data were not much good. Before either of these studies was undertaken, the commission itself had decided to obtain follow-up information on the crib-related suffocation deaths in their files. Of the 38 deaths investigated, the specific cause of death was misclassified in 18 cases. In eight cases initially reported as occurring in cribs, the babies had not been in cribs at all, but had been in beds, playpens, on a pillow, between a bed and a crib, or in a bassinet ( 4). Despite the discovery of this somewhat startling level of inaccuracy in the CPSC data, the authors of both studies built their far-reaching conclusions on these shaky foundations. Responding to this criticism, Dr.ago could only argue that the crib “subset was in no way representative of the data set,” without adducing evidence for this defense ( 12). The uncritical acceptance of coroners' death certificate diagnoses of overlying is even more troubling. The authors of the CPSC report commented on the limitations of these diagnoses, but said that they were aided by the accompanying supporting information. They quoted one certificate that said “fell asleep while nursing,” as if that proved the baby had been smothered by the mother. But the CPSC must know that determining the cause of death in infancy is often a complicated task; even the best pathologists find it difficult to differentiate suffocation from sudden infant death syndrome. As Drago commented in her rebuttal to the critique by O'Hara et al ( 5), “the distinction relies on evidence from the death scene” ( 12). In this respect, unfortunately, the CPSC data were sorely deficient. This is partly because coroners in the United States are by no means uniformly competent. They range from morticians with no training in medicine to highly expert medical pathologists who know how to look for the many causes of death in infancy. A review of CPSC death certificates showed that overlying is a common diagnosis in some areas of the country but rare in others. In an attempt to clarify this clustering of diagnoses, O'Hara carried out telephone interviews with coroners in places from which many cases of overlying were reported. Some of these officials actually said they would not diagnose sudden infant death syndrome if a baby had been sleeping with an adult. They apparently assumed that the otherwise unexplained death of a co-sleeping infant must be due to overlying ( 5). In short, a coroner's statement on a death certificate in these cases cannot be assumed to be accurate without additional evidence, and this necessary step was omitted by the authors of the CPSC report. Another complicating issue is the risk posed by a bed-sharing adult who is under the influence of alcohol and drugs. An earlier study had drawn attention to substance abuse as a possible contributing factor in overlying deaths ( 13). In an article published after the CPSC data-based studies, an English investigation of the relationship between sudden infant death syndrome and co-sleeping showed a clear increase (3.4 odds ratio) for recent maternal alcohol use ( 14,15). However, in only two of the CPSC cases was drug abuse reported. The CPSC authors noted that only “limited information” is provided on death certificates, but they made no further attempt to find out if drugs or alcohol played a role in the other cases. Like all the rest of the mammals, human infants and human mothers have slept together from the beginning of the species. It is convenient, and it keeps the babies warm. In cool climates really no alternative to bed-sharing was available until the invention of adequate home heating a few hundred years ago. In addition, until about the last century or two only the very well-to-do could afford the luxury of separate bedrooms. We know that bed-sharing powerfully increases the likelihood of successful and prolonged breastfeeding. More breastfeeding means better physical health and mental ability, both in childhood and later in adult life, and it probably means fewer infant deaths from sudden infant death syndrome ( 16–18). The emotional advantages of bed-sharing are less clear-cut. Little children certainly seem to enjoy the company of their parents in bed. Among adults, opinions regarding bed-sharing are decidedly mixed, although the devotees of the “family bed” are enthusiastic about its advantages. It is worth keeping in mind that these two changes in human sleeping patterns—sleeping in separate beds and sleeping in separate bedrooms—are brand new in terms of human history. We tend to think of them as “normal” but they are truly experimental; we have no idea what importance, if any, they may prove to have. The distinction between bed-sharing and room-sharing is important. Both the English study ( 14) just cited and a study in New Zealand ( 18) showed that babies who sleep in the same room with adults have a lower incidence of sudden infant death syndrome than babies who sleep in separate rooms. Sharing a bed is an entirely different matter. In fact, various studies have shown that an infant sleeping in the same bed as a parent who smokes has an increased risk of sudden infant death syndrome ( 19). In addition to parents who smoke or use alcohol or other sedating drugs, other risks to co-sleeping infants occur from improper bedding (comforters or pillows), soft mattresses, sofas, maternal exhaustion, and overcrowded housing conditions. Previously mostly anecdotal information was available about the extent of these hazards, but the English study has now made it possible to arrive at somewhat more definite conclusions. Blair et al surveyed 5 regions in England from 1993 to 1996, studying 325 cases of sudden infant death syndrome and comparing them with 1300 control infants ( 14). Every case was examined in detail. Research interviewers collected medical record information and met with the families within days of each death and again 2 weeks later. The cause of death was established by a multidisciplinary team after a pediatric necropsy using a standard protocol. No attempt was made to differentiate overlying deaths from sudden infant death syndrome in general; the focus was on the relationship of co-sleeping (including both bed-sharing and room-sharing) and sudden infant death syndrome. The conclusions were remarkably clear-cut: bed-sharing is least hazardous for infants over 14 weeks of age; bed-sharing is dangerous on a sofa, when either parent is a smoker, when the mother drinks alcohol, when the mother is sleep-deprived (less than 4 hours sleep during the preceding 24 hours), when a comforter (“duvet”) is used, and in overcrowded homes (more than 2 people per room in the dwelling). In the absence of these conditions, “there is no evidence that bed-sharing is hazardous…” They found that room-sharing was associated with a lower risk of sudden infant death syndrome than bed-sharing, but commented that “… further research is required to investigate whether room sharing is protective in itself or merely a marker for hidden confounders not measured in this study.” The articles based on CPSC data examined two aspects of using adult beds for babies: first, the dangers of the beds themselves, and second, the dangers of co-sleeping of adults and infants. About the problems of using adult beds for infants there has been widespread agreement for years, and the CPSC studies helped redirect attention to those aspects of adult beds that are hazardous. The risk of co-sleeping was confused rather than being clarified, however, because the limited CPSC data did not allow the authors to distinguish with any certainty deaths due to overlying from cases of sudden infant death syndrome where the baby was found dead next to a parent. I believe that their conclusions went beyond their data. No last words can be written about this set of issues, but as professionals concerned with the lives and health of families, it is necessary for us to draw at least preliminary conclusions. Here are mine. First: Beds for adults are unsafe for babies if the mattress or the bed can trap an infant in any way, or if the bedding is soft, like a feather bed, a fleece, or a soft waterbed. Parents who share their beds with their babies should make sure that the surface is firm, and that no spaces are present between the mattress and the wall or the bed frame where the baby could be entrapped. Pillows and puffy comforters or blankets that could block the baby's air supply should not be used. Second: Cribs for babies can be unsafe for exactly the same reasons as adult beds. The current CPSC standards for cribs have increased the safety of modern, properly assembled cribs with tightly fitting mattresses, but there are all too many old, faulty, damaged cribs with ill-fitting mattresses still in use. Third: Tobacco-smoking adults are dangerous for everyone around them. In addition to the fires caused by cigarettes, all the general health hazards of second-hand smoke are present (increased ear infections, respiratory infections, asthma, heart disease, lung cancer, and obesity, just to list a few). As if that were not enough, infants in the households of smokers have an increased risk of sudden infant death syndrome. The most dangerous adult is a mother who smokes and shares her bed with her baby. Fourth: The true extent of the risk of a bed-sharing adult or child overlying a baby is unknown, but it is certainly a rare cause of death in infancy. If the CPSC figures are correct; about 15 cases occur every year in the United States in contrast to the nearly 3000 cases per year of sudden infant death syndrome. Fifth: Sleeping in the same room with nonsmoking parents appears to decrease the infant's risk of sudden infant death syndrome. Sixth: Breastfeeding is more likely to be successful and prolonged when babies sleep in the same bed with their mothers. This health-enhancing effect of bed-sharing in infancy is part of the full picture that must be considered. Seventh: A separate crib for the baby in the same room as the parents' bed has some of the advantages of bed-sharing. If the adult bed cannot be made sufficiently safe, and the bedroom is large enough to accommodate a crib, this may be a comfortable compromise. Eighth: Much more information is needed. The first requirement is a system providing more reliable and complete studies of each infant death. The English study ( 14) is a model that needs to be replicated in the United States; only then can the risks of bed-sharing be properly understood. Until that time, parents and professionals alike will continue to make decisions about bed-sharing without definitive evidence. Meanwhile, the benefits of the ancient practice of bed-sharing should not be lightly abandoned on the basis of these two flawed studies.