Sudden infant death syndrome (SIDS) is the leading cause of death among infants between 1 month and 1 year of age in the developed world. In the United States, SIDS accounts for 22% of all postneonatal deaths.1 Many countries have launched educational campaigns in an effort to prevent SIDS, focusing on the modifiable factors that had been shown previously to be associated with SIDS.2 Although campaigns vary in the content of their other messages, the leading message of every campaign has been avoidance of the prone position for sleeping infants. Dramatic declines in SIDS have been attributed to these campaigns, which were found to be primarily a result of decreases in prone-sleeping rates.3–9 Other recommendations, such as those advising mothers not to smoke during pregnancy, were followed less widely.3,7
The majority of campaigns were initiated in the late 1980s to early 1990s. In the United States, the American Academy of Pediatrics recommended the nonprone sleeping position for infants in 1992.10 The national Back to Sleep campaign began 2 years later and emphasized the same recommendation.11 In 2000, on the basis of new epidemiologic evidence, the American Academy of Pediatrics recommended back sleeping as the preferred position.7 In the United States, prone-sleeping rates have declined 81%, from 70% in 1992 to 13% in 2005, and supine-sleeping rates have increased from 13% to 72% over that period.12 The US rate of SIDS has declined 55% (from 1.20 in 1000 live births in 199213 to 0.54 in 1000 live births in 2005).14
Previously, SIDS rates were found to differ by country, as had reductions in rates.15 The purpose of this report is to compare more-recent international data on SIDS and postneonatal infant mortality rates to examine trends, discuss possible reasons for differences, and provide recommendations to further reduce sudden infant deaths and to improve consistency in classifying and reporting these deaths across countries. Because assignment of cause of death may vary within and across countries in cases of sudden unexpected infant death (SUID), the postneonatal mortality (PNM) rate may be a better indicator of trends in both SIDS and other SUID for comparative purposes, because most of these deaths occur during this period of infancy.
National SIDS and PNM rates were collected from a number of sources: Web sites from national vital statistics centers and correspondence with statisticians provided on those Web sites; published reports; and correspondence with leading SIDS researchers in several countries, who, in turn, collected the data from their respective national registries or vital statistics offices. Numerous attempts were made to contact individuals if the data were not available via accessible Web sites in English or from publications. Data for SIDS and PNM rates were requested annually from 1990 to the most recent year available. The year that each country's SIDS risk-reduction campaign began was also identified. Because the ages of inclusion for the SIDS diagnosis vary by country, this information was also collected. Percentage declines in SIDS and PNM rates were calculated for the period 1990–2005.
SIDS data were obtained for 13 of 17 countries queried (Table 1). For most of these countries, there has been a striking decrease in the SIDS rate from 1990 to the most recent year available (2005 for all but Canada, for which 2004 data are the most recent). These decreases range from 40% in Argentina to 83% in Ireland. The highest SIDS rates in 1990 (≥2.0 in 1000 live births) were in Ireland, New Zealand, and Scotland. More recently, the highest SIDS rates (≥0.5 in 1000 live births) were in New Zealand and the United States. The lowest rates (≤ 0.2 in 1000) were in Japan and the Netherlands. Since 2000, the SIDS rates in most of the countries have declined minimally.
PNM rates were compared with SIDS rates in 13 countries for the period 1990–2005 (Table 1). Declines in PNM rates have occurred for all the countries, with the smallest decline in Japan (30%) and the largest decline in the Netherlands and Norway (82%). Generally, the relative declines in PNM rates were smaller than those for the SIDS rates in each respective country. The highest PNM rate was in Argentina, and the lowest was in the Netherlands.
Figure 1 shows trends in PNM rates and SIDS for each country from 1990 to 2005. For most countries, the decline in SIDS was mirrored by a decline in PNM rate, again with higher rates of decline occurring earlier in the campaigns. This provides evidence that the SIDS declines have been real. However, in later time periods, different patterns emerged. For some countries, the declines in SIDS and PNM rates have stabilized similarly (eg, the Netherlands and Sweden). For some countries, the rates of both have continued to decline gradually (eg, New Zealand), and for others the SIDS rates have stabilized, whereas the PNM rates have gone up slightly (eg, the United States).
SIDS rates were compared from 1990 to the year in which the most recent data were available for each country. The year 1990 was chosen as the baseline because SIDS risk-reduction campaigns had either just begun around that time or had not yet begun. It is evident that these campaigns have had a major impact on reducing SIDS rates.3–5,8,9,16–23 These reductions were, for most countries, well over 50%. The largest declines generally occurred in the first few years after initiation of national campaigns. Declines were also found in PNM rates in most countries. This would be expected, because the majority of SIDS deaths occur in the postneonatal period. Decreasing PNM rates occurring with decreasing SIDS rates support the decline in SIDS rates being real (especially the early declines) rather than merely the result of classifying sudden deaths differently more recently. The magnitude of decline varied and was generally smaller than the respective decline seen in the SIDS rate. Again, this would be expected, because SIDS, while being the leading cause of infant death in the postneonatal period, accounts for only a portion of the total (eg, 22% in the United States).1
Rates of SIDS, however, varied considerably across countries. In 2005, rates varied from 0.10 in 1000 live births to 0.80 in 1000. There are several possible explanations for these differences. First, the age of inclusion for SIDS differs across countries. Some countries (Canada, England and Wales, Germany, and Scotland) define SIDS as occurring from 1 week to 1 year. Others use the range of birth to 1 year (Argentina, Australia, Ireland, Japan, Sweden, and the United States) or birth to >1 year (Netherlands, New Zealand, and Norway). Although this may account for some of the differences found in the SIDS rates, it is likely to be a small effect because the number of SIDS deaths occurring in the first week of life and after 1 year are very small.24–26 The data presented in Table 1 and Fig 1 for Norway are restricted to age <1 year.
Second, as a disorder that is diagnosed by excluding other identifiable causes of death, different definitions of SIDS also contributed to the variation in rates seen not only across countries but within countries. In a recent study by Byard and Marshall,27 50 articles published in 2005 were reviewed in which the validity of the conclusions depended on accurately defining SIDS. One of 5 definitions were searched for in each article: the 1969 Seattle definition,28 the National Institute of Child Health and Human Development definition,29 the San Diego definition,30 those that used a nonstandard definition, and those for which no definition was provided. The authors found that 58% of the articles had used a nonstandard definition or provided no definition. The study highlighted the need for standard definitions to ensure validity of research data and comparability of data across centers.
Third, the content and use of death-scene investigation and autopsy protocols in cases of SUID vary also across different locales. Although standard protocols have been recommended internationally and nationally,31–33 local conditions determine what and how extensively they are applied. For example, autopsy rates are lower in the Netherlands34 and Japan35 compared with other countries. Without a thorough autopsy, investigation of the location of death, and review of pertinent medical history, it is difficult to distinguish between SIDS and other causes of SUID.
Even with similar definitions and protocols, there can be large differences in assignment of cause of death, with some pathologists underdiagnosing SIDS and others applying the diagnosis too liberally.36 In addition, there is evidence that with implementation of more comprehensive autopsy and scene-investigation protocols, there have been shifts in the classification of the cause of death within countries. For example, this “diagnostic shift” has been found in South Australia over the period 1994–1998. Mitchell et al examined 114 consecutive cases of sudden unexpected death.37 The number of SIDS cases declined from 1989–1993 to 1994–1998, whereas there was an increase in deaths attributed to accidents or classified as “undetermined.”1 The authors concluded that, although there was a real decline in SIDS, there was also a diagnostic transfer. Scene-investigation reconstructions increased the number of asphyxia deaths (“sleeping accidents”), and more social background information and subtle nonlethal injuries increased the number of undetermined deaths. Similarly, the use of comprehensive autopsy and scene-investigation protocols, along with adhering to a strict definition of SIDS,29 in the Chicago Infant Mortality Study resulted in a high rate of the undetermined diagnosis.25 An analysis by Malloy and MacDorman1 of national US data suggested, too, that a diagnostic shift is occurring across the country. They demonstrated that over the period 1999–2001, the PNM rate did not change and the postneonatal SIDS rate declined slightly. Most (90%) of this decline in SIDS was accounted for by an increase in the rates of unknown/unspecified deaths and suffocations (in bed or other). These findings were confirmed by Shapiro-Mendoza et al33 for all infant deaths from birth to 1 year for the years 1989–2001. Using linked birth/death certificate data, they found that from 1999 to 2001, the decline in SIDS rates was offset by increased rates of cause unknown/unspecified and accidental suffocation and strangulation in bed. Infant mortality rates attributable to the other causes of sudden unexpected death (ie, other accidental suffocation and strangulation, neglect, abandonment, and other maltreatment syndromes) remained unchanged. The authors also examined risk factors for SIDS and the other causes of SUID for the various time periods examined and found that the risk factors remained stable over time and were common to all the SUID infants, suggesting that the decline in SIDS from 1999 to 2001 was not likely a true decline but, rather, related to the way in which these infant deaths were classified.
Similarly, an investigation in England found that declines in the SIDS rate have been accompanied by increases in the rate of unascertained deaths.38 The author concluded that analyses of unexplained infant deaths should include both categories and posited that these trends could be the result of changes in the certification practices of coroners. This theory was borne out by a study conducted by Limerick and Bacon39 of pathologists in England who performed infant autopsies in cases of sudden infant death. They found wide variations in the pathologists' use of the terms “SIDS” and “unascertained.” Use of the latter was common when infants were sharing a bed with an adult or when suspicious features were present.
Shields et al40 conducted a study of infant deaths handled by the medical examiners' offices in Kentucky and attributed to SIDS, positional asphyxia, overlay, and undetermined cause and manner of death. They compared these categories before and after instituting a standard classification scheme that was based on, but not identical to, the San Diego classification.30 They found that there was a greater number of deaths deemed undetermined and a smaller number of SIDS deaths before the policy change compared with the reverse after this change. These types of classification variations are likely occurring in other countries, where analyses by cause of death need to be conducted to fully understand respective trends. Thus, for more recent periods, the “true” decline in SIDS rates in some countries may be lower than the statistics would imply.
Might there be another explanation for the slightly increased PNM rate in the United States more recently (eg, an increase in low birth weight infants being kept alive into infancy)? This is unlikely, because the infant mortality rate among low birth weight infants has steadily declined,41 even as the rates of preterm and low birth weight infants have increased.42
Finally, risk factors for SIDS vary across countries and, therefore, are likely to contribute to the variability in rates. For example, smoking rates are high among the Maori in New Zealand and American Indians, groups in which the rate of SIDS remains high.41,43 Infant prone-sleeping rates have declined to single digits in several countries (eg, in Western Sweden the rate was 5.6% in 2003–2004)44 while remaining higher in others (eg, in the United States prone placement was 12.2% in 2006 for all infants and 21.9% for black infants).12
Introduction of the 10th revision of the International Classification of Diseases (ICD-10) in 1999 could have influenced some of the changes in SIDS rates; the ICD-9 was used from 1979 to 1998. However, the effect of the new revision is likely to be insignificant. Malloy and MacDorman* examined the possible effect of different ICD revisions on trends in cause-specific mortality rates, adjusting the rates for the major causes of SUID for the period 1992–1998 under ICD-9 to be comparable to ICD-10 rates. The adjusted rates were not significantly different from the unadjusted rates.
CONCLUSIONS AND RECOMMENDATIONS
There have been impressive reductions in SIDS deaths around the world. These declines seem to be real and attributable, in large measure, to risk-reduction activities, especially placing infants supine to sleep.4,5,8,19 However, rates have stabilized in the majority of countries, and in some countries they remain unacceptably high, which highlights the need for risk-reduction activities to be continued, especially in communities with the greatest burden of SIDS. In the United States in 2004, the National Infant Sleep Position Survey revealed that 12.9% of infants were placed prone for sleep.12 Using the pooled odds ratios from 7 case-control studies conducted in the post–Back to Sleep period for prone sleeping position (6.02), we estimate that 47% of the annual SIDS deaths in the United States could be attributed to prone sleeping (Appendix).45 Using a more conservative estimate of an odds ratio of 2.5 (based on 2 studies in the United States),25,46 an estimated 27% of SIDS deaths could be attributed to placing infants prone to sleep (Appendix). Thus, working toward eliminating the use of this position has the potential to reduce the number of SIDS deaths in the United States significantly.
Differences in rates and trends are also influenced by diagnostic shifts that have occurred. Consequently, several classifications for SIDS and SUID have been proposed as a way to achieve greater accuracy and consistency in diagnosis within and across countries.30,36,47,48 The Nordic countries have been successful in adopting standard criteria to diminish previously identified discrepancies in SIDS rates.49–51 It is essential that more widespread consensus on the definition and classification of sudden unexpected death in infancy be achieved so that national and international comparisons are more meaningful. Several initiatives in the United States are underway to develop a standardized approach, including development by the Centers for Disease Control and Prevention of a national surveillance system for SUIDs that would collect data on modifiable risk factors for SIDS and SUID and on the conduct and quality of the death-scene investigation and autopsy; a standardized classification scheme for cause of death will also be developed.33 In addition, legislation is being considered that would fund this and other related activities, including training for those who investigate infant deaths and certify cause of death.
The collection of SIDS and other infant mortality data internationally needs to be easier. Although using the Internet has provided better access to data, in most cases the Web sites were inadequate to achieve the goals for our comparisons. The reasons for this include: data were located in several different reports with sometimes conflicting numbers; not all years were included, or years were combined; and most Web sites were not in English. In 2 countries, the most recent data were not yet available on their vital statistics Web sites, and a cost was involved in 1 case to obtain these data. Without the collaboration of SIDS researchers and vital-records staff, this project would not have been possible. In addition, the results presented in this article are limited to the countries from which data were available and, thus, do not provide a fully representative profile of international SIDS and PNM rates. In the developing world where resources are severely limited, autopsies and scene investigations are not routinely performed, and other causes of infant mortality, such as infectious diseases, predominate.52 Studies and methodologies need to be developed to elucidate the extent of SIDS and SUID within these less-developed countries.
Given the challenges outlined above, a first step would be for countries to report annual statistics on the number and rate of SIDS deaths; the number of live births; the number and rate of postneonatal deaths, ideally with rates for the leading causes within that category; the age range for which the SIDS diagnosis is applied; and the definition of SIDS, if 1 standard is used. Ideally, these would be available on national vital-statistics Web sites in English. International research groups and other bodies should continue to work toward developing a uniform classification of SIDS and SUID.
Finally, in addition to infant sleep positioning, other well-established risk factors should receive attention, such as maternal smoking in pregnancy, infant overheating, and soft bedding.15,53,54 These are especially important in countries that have achieved high supine sleeping rates and which have seen increases in other risk factors such as smoking among women.55,56 In addition, emerging modifiable risk factors need to be publicized and discussed with families and caregivers of young infants. Epidemiologic research should be conducted to determine if promotion of “reduce-the-risk” messages other than sleep position will lead to further declines in SIDS and SUID. In addition, research that investigates the underlying pathophysiological etiology (or etiologies) of SIDS should be ongoing, because it is likely that reaching the goal of eliminating SIDS will occur only when we fully identify its causes.
Calculation of Population Attributable Risk for Prone-Sleep Position Among Infants in the United States, 2004
2004 SIDS rate = 0.546 in 1000 live births, N = 2247, total annual live births = 4115385
Given 12.9% prevalence of prone-sleep position used for last night's sleep in newborns in 2004, we estimate that there were 530885 infants who were sleeping prone.12
Hence, we have marginal counts for the 2 × 2 table:
We are interested in estimating cell count a, which is the number of SIDS in prone-sleeping infants.
Table of No. 3 and equation of No. 4 from above give a nonlinear equation in a: .
Solving this equation we get:
For OR of 2.5, the count of a is 607; for OR of 6.0 the count of a is 1057 Therefore, the population fraction of deaths from SIDS that could be attributable to prone sleeping is 607/2247 = 27%, and 1057/2247 = 47%, respectively.
Collection of data, analysis of data, and manuscript preparation were supported by the Health Resources and Services Administration, Academic Administrative Units in Primary Care-Family Medicine (grant D12HP00148). The sponsor had no role in study design, collection, analysis, or interpretation of data, the writing of the report, or the decision to submit the article for publication.
We thank Rachel Moon and David Slawson for manuscript review and Mir Siadaty for statistical support. Appreciation is also extended to the following people for providing data for their respective country: Enrique Abeyá (Argentina), Bernt Alm (Sweden), Brooke Black (Australia), Peter Blair (England and Wales), Hazel Brooke (Scotland), Fiona Brown (Scotland), Peter Burke (Australia), Aurore Côté (Canada), Dorothy Ford (Australia), Stephanie Fukui (Japan), Monique L'Hoir (the Netherlands), Cliona McGarvey (Ireland), Tom Matthews (Ireland), Edwin Mitchell (New Zealand), Gunvar Østevold (Norway), Denise Sheard (Australia), Toshiko Sawaguchi (Japan), Barry Taylor (New Zealand), Åshild Vege (Norway and Sweden), Bregje Van Sleuwen (the Netherlands), and Mechtild Venneman (Germany). We also appreciate the assistance of Leanne Raven and SIDS and Kids of Australia for providing funds to obtain Australian data.
- Accepted June 4, 2008.
- Address correspondence to Fern R. Hauck, MD, MS, University of Virginia School of Medicine, Department of Family Medicine, PO Box 800729, Charlottesville, VA 22908-0729. E-mail:
↵* This refers to ICD category “unknown and unspecified;” it is also called “undetermined” or “unascertained.” The ICD-9 code was 799.9 and ICD-10 is R99.
The authors have indicated they have no financial relationships relevant to this article to disclose.
Opinions expressed in these commentaries are those of the author and not necessarily those of the American Academy of Pediatrics or its Committees.
- ↵Malloy MH, MacDorman M. Changes in the classification of sudden unexpected infant deaths: United States, 1992–2001. Pediatrics.2005;115 (5):1247– 1253
- ↵Creery D, Mikrogianakis A. Sudden infant death syndrome. Clin Evid.2005;(13):434– 443
- ↵Mitchell EA, Brunt JM, Everard C. Reduction in mortality from sudden infant death syndrome in New Zealand: 1986–92. Arch Dis Child.1994;70 (4):291– 294
- Wigfield R, Fleming PJ. The prevalence of risk factors for SIDS: impact of an intervention campaign. In: Rognum TO, ed. Sudden Infant Death Syndrome: New Trends in the Nineties. Oslo, Norway: Scandinavian University Press; 1995:124–28
- ↵American Academy of Pediatrics, Task Force on Infant Position and SIDS. Changing concepts of sudden infant death syndrome: implications for infant sleeping environment and sleep position. Pediatrics.2000;105 (3 pt 1):650– 656
- ↵American Academy of Pediatrics, Task Force on Infant Positioning and SIDS; Kattwinkel J, Brooks J, Myerberg D. Positioning and SIDS [published correction appears in Pediatrics. 1992;90(2 pt 1):264]. Pediatrics.1992;89 (6):1120– 1126
- ↵American Academy of Pediatrics, Task Force on Infant Position and SIDS; Kattwinkel J, Brooks J, Keenan ME, Malloy M. Infant sleep position and sudden infant death syndrome (SIDS) in the United States: joint commentary from the American Academy of Pediatrics and selected agencies of the federal government. Pediatrics.1994;93 (5):820– 820
- ↵National Infant Sleep Position. Public-access Web site. Available at: dccwww.bumc.bu.edu/ChimeNisp/Main_Nisp.asp. Accessed May 22, 2008
- ↵Hauck FR. Changing epidemiology. In: Byard RW, Krous HF, eds. Sudden Infant Death Syndrome: Problems, Progress & Possibilities. London, United Kingdom: Arnold; 2004:31–57
- ↵de Jonge GA, Burgmeijer RJ, Engelberts AC, Hoogenboezem J, Kostense PJ, Sprij AJ. Sleeping position for infants and cot death in the Netherlands 1985–91. Arch Dis Child.1993;69 (6):660– 663
- Hiley CMH, Morley CJ. Evaluation of government's campaign to reduce risk of cot death. BMJ.1994;309 (6956):703– 704
- Willinger M, Hoffman HJ, Hartford RB. Infant sleep position and risk for sudden infant death syndrome: report of meeting held January 13 and 14, 1994, National Institutes of Health, Bethesda, MD. Pediatrics.1994;93 (5):814– 819
- ↵Hauck FR, Moore CM, Herman SM, et al. The contribution of prone sleeping position to the racial disparity in sudden infant death syndrome: the Chicago Infant Mortality Study. Pediatrics.2002;110 (4):772– 780
- ↵Beckwith J. Discussion of Terminology and Definition of the Sudden Infant Death Syndrome. Seattle, WA: University of Washington Press;1970
- ↵Krous HF, Beckwith JB, Byard RW, et al. Sudden infant death syndrome and unclassified sudden infant deaths: a definitional and diagnostic approach. Pediatrics.2004;114 (1):234– 238
- ↵Krous HF, Byard RW. International standardized autopsy protocol for sudden unexpected infant death. In: Byard RW, Krous HF, eds. Sudden Infant Death Syndrome: Problems, Progress & Possibilities. London, United Kingdom: Arnold; 2001:319–333
- Centers for Disease Control and Prevention. Sudden unexplained infant death investigation report form (SUIDIRF). In: Byard RW, Krous HF, eds. Sudden Infant Death Syndrome: Problems, Progress & Possibilities. London, United Kingdom: Arnold; 2001:334–347
- ↵Shapiro-Mendoza CK, Tomashek KM, Anderson RN, Wingo J. Recent national trends in sudden, unexpected infant deaths: more evidence supporting a change in classification or reporting. Am J Epidemiol.2006;163 (8):762– 769
- ↵Corbin T. Investigation into sudden infant deaths and unascertained infant deaths in England and Wales, 1995–2003. Health Stat Q.2005;(27):17– 23
- ↵Limerick SR, Bacon CJ. Terminology used by pathologists in reporting on sudden infant deaths. J Clin Pathol.2004;57 (3):309– 311
- ↵Mitchell EA, Tuohy PG, Brunt JM, et al. Risk factors for sudden infant death syndrome following the prevention campaign in New Zealand: a prospective study. Pediatrics.1997;100 (5):835– 840
- ↵Alm B, Mollborg P, Erdes L, et al. SIDS risk factors and factors associated with prone sleeping in Sweden. Arch Dis Child.2006;91 (11):915– 919
- ↵Li DK, Petitti DB, Willinger M, et al. Infant sleeping position and the risk of sudden infant death syndrome in California, 1997–2000. Am J Epidemiol.2003;157 (5):446– 455
- Gregersen M, Rajs J, Laursen H, et al. Pathologic criteria for the Nordic study of SIDS. In: Rognum TO, ed. Sudden Infant Death Syndrome: New Trends in the Nineties. Oslo, Norway: Scandinavian University Press; 1995:50–58
- ↵Rognum TO, Arnestad M, Bajanowski T, et al. Consensus on diagnostic criteria for the exclusion of SIDS. Scand J Forens Sci.2003;3 (4):62– 73
- ↵Hunt CE, Hauck FR. Sudden infant death syndrome. CMAJ.2006;174 (13):1861– 1869
- ↵International Network of Women Against Tobacco. Fact sheet: women and smoking in Finland. Available at: www.inwat.org/pdf/finland.pdf. Accessed: January 6, 2005
- ↵World Health Organization. “Young women's smoking crisis” set to hit Asia [press release]. Available at: www.who.int/inf-pr-1999/en/pr99-69.html. Accessed June 10, 2008
- Copyright © 2008 by the American Academy of Pediatrics