 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
* Office of Planning and Evaluation, the Consumer Product Safety Commission, Bethesda, Maryland, and Loyola College, Baltimore, Maryland
Consumer Product Safety Commission, Directorate for Epidemiology, Bethesda, Maryland
Pulmonary Medicine Division, Department of Pediatrics, St Louis University School of Medicine, St Louis, Missouri
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods. We reviewed all accidental suffocation deaths among infants 
11 months of age reported to the United States Consumer Product Safety Commission from 1980 through 1983 and 1995 through 1998. We compared infants’ ages and other demographic data, the sleep location and surface used, and the reported mechanism or pattern of death. For 1995–1998, we used data on sleep location from an annual survey of randomly selected households of living infants younger than 8 months, collected as part of the National Infant Sleep Position Study at the National Institute of Child Health and Human Development, to calculate risk for death as a result of suffocation in cribs, in adult beds, and on sofas or chairs.
Methods. The number of reported suffocation deaths by location were compared between the 1980s and 1990s using logistic regression modeling to calculate odds ratios (OR), 95% confidence intervals (CI), and P values. Comparative risks for suffocation deaths on a given sleep surface for infants in the 1990s were examined by calculating rates of death per 100 000 exposed infants and comparing the 95% CI for overlap.
Results. From the 1980s, 513 cases of infant suffocation were considered; from the 1990s, 883 cases. The number of reported suffocation deaths in cribs fell from 192 to 107, the number of reported deaths in adult beds increased from 152 to 391, and the number of reported deaths on sofas or chairs increased from 33 to 110. Using cribs as the reference group and adjusting for potential confounders, the multivariate ORs showed that infant deaths in adult beds were 8.1 times more likely to be reported in the 1990s than in the 1980s (95% CI: 3.2–20.3), and infant deaths on sofas and chairs were 17.2 times more likely to be reported in the 1990s than in the 1980s (95% CI: 5.0–59.3). The sleep location of a subset of cases from the 1990s, 348 infants younger than 8 months at death, was compared with the sleep location of 4220 living infants younger than 8 months. The risk of suffocation was approximately 40 times higher for infants in adult beds compared with those in cribs. The increase in risk remained high even when overlying deaths were discounted (32 times higher) or the estimate of rates of bedsharing among living infants doubled (20 times higher).
Conclusions. Reported deaths of infants who suffocated on sleep surfaces other than those designed for infants are increasing. The most conservative estimate showed that the risk of suffocation increased by 20-fold when infants were placed to sleep in adult beds rather than in cribs. The public should be clearly informed of the attendant risks.
Key Words: sudden death • infant; suffocation • infant; consumer product safety; sleep • infant
Abbreviations: SIDS, sudden infant death syndrome • CPSC, US Consumer Product Safety Commission • NICHD, National Institute of Child Health and Human Development • e-code, external cause of death • CI, confidence interval • NCHS, National Center for Health Statistics • OR, odds ratio
No epidemiologic studies have directly compared the risk of suffocation for infants who sleep in cribs, in adult beds, and on sofas. Although the rate of death from sudden infant death syndrome (SIDS) has declined substantially as the result of campaigns to foster back sleeping in the United States and elsewhere, the proportion of sudden, unexpected deaths diagnosed as accidental suffocation and related preventable causes may be increasing.1 For example, a recent study using US- linked birth/death data found that whereas other infant deaths were decreasing, overall approximately 20% a year, deaths diagnosed as mechanical suffocation were increasing, on the average approximately 3.7% a year.2 During a similar period, Drago and Dannenberg3 found increasing trends for specific hazards leading to suffocation, such as wedging and oronasal obstruction.
Other studies have described how suffocation deaths may have occurred in various infant sleeping environments,3–5 but these studies were "numerator-based" and did not have a denominator for comparison. Thus, they did not permit calculation of the risk of sleeping in different locations such as cribs and adult beds. In this analysis, we used data describing sleep practices for 348 infants whose death was identified as suffocation in cribs and adult beds (we use the term "adult" beds because approximately 98% of the bed-related deaths outside of cribs were on beds designed for adults) from 1995 to 1998 and compared their sleeping locations with comparable data from a study of 4220 living infants from 1995 to 1998. Information about the infant deaths was obtained primarily from death certificates for deaths reported to the US Consumer Product Safety Commission (CPSC). Information about living infants was obtained from a survey of caregivers for infants conducted by the National Institute of Child Health and Human Development (NICHD) during the same period.6,7
We also compared reported infant deaths diagnosed as suffocation from 1980 through 1983 with those from 1995 through 1998 to examine any changes in infant characteristics, sleeping locations, and suffocation patterns that may have occurred over the decades. We were interested in changes over the past 20 years, although we know that reporting changes and diagnostic shifts have occurred during this period. This attempt at describing changes in victims and patterns of suffocation is worthwhile because the proportion of deaths diagnosed as accidental suffocation seems to be increasing.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Death Certificate database contains information from death certificates that CPSC obtains from all 50 states, the District of Columbia, and New York City.8 Death certificates that fall within specified external cause of death categories (e-codes) are obtained9 because they are likely to report specific consumer products. CPSC receives all suffocation e-codes, with the exception of the e-code for "suffocation by falling earth." CPSC also collects information on deaths through its Medical Examiners and Coroners Alert Project and through other sources (eg, police and fire reports, media articles). This information is stored in the Injury and Potential Injury Incidents database. The In-depth Investigations database contains information from CPSC follow-up investigations of specific injuries and deaths.
Searches of the 3 databases produced 3081 cases. After removing duplicate and out-of-scope cases (eg, fire and high-chair deaths) and limiting age to 11 months, 1396 cases were available for analysis: 513 deaths from the 1980s and 883 deaths from the 1990s. Identifying information in CPSC databases is maintained following formal procedures that ensure confidentiality.
Suffocation Patterns
Two raters (N.J.S., G.W.R.) independently classified each death as to the specific suffocation pattern based on the narrative in the report. The agreement rate was 96.4% for all sleeping locations and 98.4% for cribs, adult beds, and sofa/chairs. The criteria for classification for each specific pattern were as follows:
Information was available for a number of variables including, age, sex, race, state, e-code, and date of death. Sleeping location was identified from the narrative in the death certificate, injury and potential injury report, or in-depth investigation. Race was dichotomized into white and minority categories consistent with accepted practice10; within the minority category, blacks represented approximately 92.3% of the total. Season of the year was categorized as winter (for the months October to March) and summer (April to September).
Living Infants
Information was obtained from the 1990s for both living and deceased infants who died in the 1990s, but we are unaware of comparable information for living infants for the 1980s. Thus, our risk analysis is restricted to the 1990s.
Information on the sleep practices for a cohort of living infants was obtained from the NICHD’s National Infant Sleep Position Study,6,7 a survey conducted annually from 1994 to 1998. Telephone interviews were completed by nighttime caregivers of infants younger than 8 months from a randomly selected annual sample of >1000 households (range: 1043–1050 households per year). We used data from this survey on sleeping location for 1995–1998, the years that coincided with the deaths in our study. Infants’ sleeping locations were identified by responses to the question, "Where did your infant usually sleep during the past 2 weeks?" Details of the methods used in the study have been reported elsewhere.6,7
Statistical Analyses
All analyses, with the exception of risk, were calculated for infants 11 months old. For risk calculations, we used infants who died of suffocation at younger than 8 months during 1995–1998 to be consistent with the sample from the National Infant Sleep Position Study.
For univariate analyses, 
2 was used to test for differences in relative frequency distributions, with the Fisher exact test or Yates corrected 2 used for 2 x 2 tables. Differences among means were tested using analysis of variance. Logistic regression modeling was used to estimate univariate odds ratios (ORs), 95% confidence intervals (CIs) and P values for comparisons between the 1980s and 1990s. Trends in suffocation rates over time were tested using Poisson regression analysis.
For the 1990s cohort, we calculated the risk of sleeping in adult beds or cribs by separate ratios using the number of deaths in cribs or beds as the numerator and an appropriate exposure measure as the denominator, along with 95% CIs. Exposure, the number of living infants who slept in cribs or beds, was based on the NICHD data showing the distribution of various sleeping locations for 1995–1998 and the number of living infants from census data for 1995–1998.11 CIs for rates based on <100 deaths were computed using the Poisson distribution; otherwise, CIs were computed using the binomial distribution. The difference between 2 rates from these distributions is considered to be statistically significant when there is no overlap between the 95% CIs.11
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
2 analysis, P < .0001); an increase in reports of suffocation deaths in adult beds, from 30% in the 1980s to 44% in the 1990s (2 analysis, P < .0001); and an increase in reports of suffocation deaths on sofas or chairs, from 6% in the 1980s to 12% in the 1990s (2 analysis, P < .0001).
|
The interaction of age of the infant with adult beds and the interaction of age with sofa or chairs in the multivariate model described above both were significant. Figures 1 and 2 show that this change was primarily attributable to the increase in the number of reported deaths in the first month of life for infants sleeping in adult beds and sofas or chairs. For these 2 sleeping locations, more infants in the 1990s were reported to have died of suffocation during the first month of life (25.5%) than any other month, compared with 13.2% who died during the first month of life in the 1980s.
|
|
2 analyses, P 
.23). Table 3 shows the suffocation patterns for the 3 most common sleeping locations where suffocation occurred.
|
Although there were fewer crib-related entrapment deaths in the 1990s, approximately 72% of the reports noted that the entrapments were the result of a failure of the crib’s structure or hardware (eg, entrapment from loose or broken slats, broken crib rails, loose screws); a mattress that was too small; an old crib with slats that were too wide; or in a few cases, a crib that was misassembled.
Suffocation Deaths From Adult Bed Hazards
The primary suffocation pattern associated with deaths in adult beds was entrapment in both decades (Table 3). In the 1980s, there were 89 reports of entrapments in adult beds compared with 188 deaths in the 1990s. Most of the deaths in both decades occurred from entrapment between the bed and the wall (52.3%), or the infant was wedged (29.6%) between the headboard or footboard and mattress or between the bedrail and another bed component.
The diagnosis by medical examiners and coroners that overlying of an infant while sharing an adult bed was the "cause of death" remains controversial. More overlying deaths were reported by medical examiners and coroners in the 1990s (70 deaths) than in the 1980s (7 deaths). In approximately 40.3% of the cases (31 of 77), the narratives reported that a third party found the infant covered by an adult or a child, there were compression marks on the infant, or other findings suggesting the likelihood of overlying (eg, infant sleeping in twin bed with 2 adults). In both decades, overlying deaths were associated with very young infants, with an average age of 1.9 months. Only 1 overlying death occurred after 6 months of age, a report of a 10-month-old found with another child over him.
Suffocation Deaths From Sofa/Chair-Related Hazards
The primary suffocation patterns associated with deaths on sofas and chairs were entrapment and overlying (Table 3). Entrapment deaths occurred in 11 of 33 reported in the 1980s, and 31 of 110 reported in the 1990s. For deaths in sofas/chairs, overlying reported for 3 of 33 in the 1980s, and 33 of 110 in the 1990s. The average age of these infants reported to have suffocated from entrapment or overlying in sofas or chairs in the 1990s was 2.8 months. Sixty percent (66 of 110) were 2 months old or younger.
There were 7 reports of nonspecific suffocations on sofas or chairs in the 1980s and 33 deaths in the 1990s (Table 3). Reports from the 1990s show that 60.6% of the deaths on sofa and chairs classified as nonspecific suffocations also reported that the infant was sharing a sofa or a chair with another person, usually an adult. The average age for these deaths was 3.2 months.
Comparison of Risk in Cribs With Risk in Adult Beds
As noted above, we compared the risk of suffocation in cribs and adult beds by restricting the deaths to infants younger than 8 months to be comparable with the NICHD study of the sleep practices of living infants younger than 8 months. Both data sets pertain to the years 1995–1998. Comparison of risk was calculated by comparing deaths of infants in cribs or adult beds with the NICHD estimate of infants usually sleeping in cribs or adult beds (Table 4). For deaths from CPSC databases for infants younger than 8 months, there were a total of 68 suffocation deaths in cribs and 348 suffocation deaths in beds during the 4-year period. During that same time period, NICHD data shows that approximately 70% of infants usually slept in cribs, whereas 9% usually slept in beds. Applying the percentages from the NICHD study to the number of live births, we estimate that in the 1990s, on average, 2.6 million infants usually slept in cribs and 340 000 infants usually slept in adult beds.
|
We conducted 2 sensitivity analyses to examine the effect on the risk calculations for adult beds. The first analysis restricted the cases to nonoverlying deaths, resulting in a decreased numerator for the risk calculations. The second analysis doubled the estimated number of living infants sleeping in adult beds, resulting in an increased denominator for the risk calculations (Table 5).
|
For the second sensitivity analysis, we doubled the estimated number of living infants sleeping in adult beds because of concerns that bed sharing is underreported. This increased the denominator from 9.0% to 18.0% of living infants sleeping in adult beds for the 4-year period. The adjusted risk was approximately 12.7 deaths per 100 000 infants younger than 8 months from 1995 to 1998 and again remained significantly and substantively higher than the risk of death in cribs.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The increase in the number of reported suffocation deaths from the 1980s to the 1990s involving adult beds may have several explanations, including increased reporting and diagnostic shift, in addition to the possibility of a real increase in the number of suffocation deaths. One obvious example is that suffocation hazards linked to soft bedding were not well known in the 1980s, and thus a reported increase in deaths linked to soft bedding may be partially attributable to underreporting that occurred in the 1980s. Thus, it is likely that some deaths in adult beds, which may have been diagnosed as SIDS until the early or mid-1990s, are now being designated as accidental suffocation.12,13 This diagnostic shifting is likely attributable to a new appreciation of how infants might subtly suffocate14 and to greater reliance on death scene information to document the occurrence of accidental deaths that were once considered unexplained and thus attributed to SIDS.12 The net result of these 2 changes in the understanding and investigation of sudden, unexpected infant death may be an increase in the reporting of suffocation deaths, particularly among younger infants sleeping in adult beds and other noninfant sleep surfaces.
Younger infants may be at greater risk in adult beds15 (Figs 1 and 2, Table 4) because they lack motor skills to escape potential threats within the sleep environment (Table 3, both decades).16,17 Also, in a recent case-comparison study in Cleveland, Ohio, from 1992 to 1996, Carroll-Pankhorst et al18 found that infants who died of SIDS while bedsharing were significantly younger than those who died while sleeping alone (9.1 weeks vs 12.7 weeks). These investigators speculated that some infant deaths diagnosed as SIDS were actually suffocation and that "younger, less vigorous infants" may be more susceptible. Our results are consistent with these findings from Cleveland.
In large epidemiologic studies of SIDS, it seems that sleeping in the parents’ room without bedsharing confers the least risk, compared with sharing a sleep surface (most risk) and sleeping in another room (intermediate risk).19 There are obvious advantages to recommending cribs, bassinets, and other infant products that are designed for use by infants. For beds not designed for infants only, it is difficult to control potential hazardous arrangements causing suffocation, with or without entrapment. In contrast, relatively straightforward safety standards for cribs have been used successfully since 1973 to reduce crib-related suffocation deaths. Mandatory standards developed during the 1970s include requirements for side height, slat spacing, and mattress fit.20,21 Voluntary standards developed since 1985 include requirements to prevent entanglement on corner posts and collapse of the crib.22–24
The question regarding the usual sleeping location of living infants from the NICHD study may not have identified infants who sometimes or occasionally slept on an adult bed. For this reason, the results may understate the exposure to adult beds slightly. Given that the difference between the risks for adult beds and cribs is approximately 40:1, it is unlikely that this would make much difference in the conclusion that adult beds have a higher risk for infants than cribs. The published data suggest a probable increase in risk associated with bed sharing for the previous night compared with usual sleep,25,26 suggesting that the risk may be underestimated when a question about bedsharing for "usual sleep" is asked.
There are several limitations of the data sets used in this study. CPSC receives most but not all of the deaths reported to NCHS, and therefore the deaths in this study should be considered a minimum number. A study of the completeness of CPSC death data for all e-codes found that CPSC received an average of 82% of the number of cases in the NCHS mortality file.27 Data extracted from death certificates also do not provide information on other potentially important factors, such as the position of the infant, or whether a death-scene investigation was completed. However, Malloy28 noted that this lack of information about the completion of a death scene may "allow for an overestimation of the number of SIDS deaths that are recorded and an underestimation of other related causes of death, such as asphyxia and accidental suffocation." Despite these limitations, the data used in this study provide a comprehensive view of suffocation deaths in the United States for 2 time periods. There is no alternative database in the United States that captures product-related deaths at the national level.
Although the deaths reported to CPSC are a minimum number of known deaths, the number of deaths in cribs is likely to be a fairly complete count of all crib-related deaths. The CPSC actively seeks information on all crib-related deaths and deaths related to other children’s products from death certificates, through medical examiners and coroners, and from other sources. These data are monitored daily. For this reason, the suffocation data for cribs and other products intended for use by children may be more complete than the data for adult beds and other products not designed for children. We also believe that the number of deaths in adult beds or on sofas or chairs is likely to be an undercount, because for many overlying deaths, the sleeping location was "not reported" (Table 2) and thus may have involved deaths on adult beds or sofas and chairs for which more specific data were lacking. Thus, we speculate that our risk calculations (Tables 4 and 5) include fairly complete data for suffocation deaths in cribs, whereas the number of suffocation deaths in adult beds may be undercounted.
Infant deaths diagnosed as suffocation in adult beds and on sofas or chairs are being increasingly reported in the United States, whereas suffocation deaths in cribs are declining. This may be attributable to a diagnostic shift or increased reporting of adult bed and sofa or chair suffocations, or there may be a real increase in the number of these deaths. Whatever the explanation, these data indicate that there are more infants dying in adult beds and on sofas or chairs than we were aware of in the 1980s. Suggestions from some29 that the risk associated with falling and entrapment might be lessened by, for example, pushing an adult bed near the wall are of unproven efficacy, have been known to result in infant deaths,3 and should be discouraged.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Reprint requests to (J.S.K.) Pulmonary Division, Department of Pediatrics, St Louis University School of Medicine, 1465 South Grand Blvd, St Louis, MO 63104-1095. E-mail: kempj{at}slu.edu
The opinions expressed in this article are those of the authors in their private capacities (N.J.S., G.W.R.) and do not necessarily represent the views of the Consumer Product Safety Commission.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1508)1509)This article has been cited by other articles:
 |
|
 |
|
  A. Jackson and R. Y. Moon An Analysis of Deaths in Portable Cribs and Playpens: What Can Be Learned? Clinical Pediatrics, April 1, 2008; 47(3): 261 - 266. [Abstract] [PDF]
|
|
|
|
 |
|
 K. Brown Kirschman and G. A Smith Resale of recalled children's products online: an examination of the world's largest yard sale Inj. Prev., August 1, 2007; 13(4): 228 - 231. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Pressley, B. Barlow, T. Kendig, and R. Paneth-Pollak Twenty-Year Trends in Fatal Injuries to Very Young Children: The Persistence of Racial Disparities Pediatrics, April 1, 2007; 119(4): e875 - e884. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S Y Pan, A-M Ugnat, R Semenciw, M Desmeules, Y Mao, and M MacLeod Trends in childhood injury mortality in Canada, 1979-2002. Inj. Prev., June 1, 2006; 12(3): 155 - 160. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C Jenny and R Isaac The relation between child death and child maltreatment. Arch. Dis. Child., March 1, 2006; 91(3): 265 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Task Force on Sudden Infant Death Syndrome The Changing Concept of Sudden Infant Death Syndrome: Diagnostic Coding Shifts, Controversies Regarding the Sleeping Environment, and New Variables to Consider in Reducing Risk Pediatrics, November 1, 2005; 116(5): 1245 - 1255. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. Lahr, K. D. Rosenberg, and J. A. Lapidus Bedsharing and Maternal Smoking in a Population-Based Survey of New Mothers Pediatrics, October 1, 2005; 116(4): e530 - e542. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. P. Kiernan and R. C. Beckerman Is It Sudden Infant Death Syndrome or Sudden Unexpected Infant Death? Pediatrics, September 1, 2005; 116(3): 800 - 801. [Full Text] [PDF]
|
|
|
|
|
|
L. B. E. Shields, D. M. Hunsaker, S. Muldoon, T. S. Corey, and B. S. Spivack Risk Factors Associated With Sudden Unexplained Infant Death: A Prospective Study of Infant Care Practices in Kentucky Pediatrics, July 1, 2005; 116(1): e13 - e20. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Vemulapalli, K. Grady, and J. S. Kemp Use of Safe Cribs and Bedroom Size Among African American Infants With a High Rate of Bed Sharing Arch Pediatr Adolesc Med, March 1, 2004; 158(3): 286 - 289. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Let Babies Sleep in Their Cribs Journal Watch Women's Health, December 30, 2003; 2003(1230): 2 - 2. [Full Text]
|
|
Read all P3Rs
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2003 American Academy of Pediatrics. All rights reserved. |
||
|
||
