The Role of Breastfeeding in Racial and Ethnic Disparities in Sudden Unexpected Infant Death

A Population-Based Study of 13 Million Infants in the United States

Melissa Bartick; Alexis Woods Barr; Lori Feldman-Winter; Mònica Guxens; Henning Tiemeier


Am J Epidemiol. 2022;191(7):1190-1201. 

In This Article

Abstract and Introduction


Sudden unexpected infant death (SUID) disproportionately affects non-Hispanic Black (NHB) and American Indian/Alaskan Native infants, who have lower rates of breastfeeding than other groups. Using 13,077,880 live-birth certificates and 11,942 linked SUID death certificates from 2015 through 2018, we calculated odds ratios and adjusted risk differences of SUID in infants who were not breastfed across 5 racial/ethnic strata in the United States. We analyzed mediation by not breastfeeding in the race/ethnicity–SUID association. The overall SUID rate was 0.91 per 1,000 live births. NHB and American Indian/Alaskan Native infants had the highest disparity in SUID relative to non-Hispanic White infants. Overall, not breastfeeding was associated with SUID (adjusted odds ratio (aOR), 1.14; 95% confidence interval (CI): 1.10, 1.19), and the adjusted risk difference was 0.12 per 1,000 live births. The aOR of not breastfeeding for SUID was 1.07 (95% CI: 1.00, 1.14) in NHB infants and 1.29 (95% CI: 1.14, 1.46) in Hispanic infants. Breastfeeding minimally explained the higher SUID risk in NHB infants (2.3% mediated) and the lower risk in Hispanic infants (2.1% mediated) relative to non-Hispanic White infants. Competing risks likely explain the lower aOR seen in NHB infants of not breastfeeding on SUID, suggesting that social or structural determinants must be addressed to reduce racial disparities in SUID.


Sudden unexpected infant death (SUID) remains the leading cause of post-neonatal death in the United States,[1] affecting approximately 3,400 infants annually.[2] According to the Centers for Disease Control and Prevention, SUID comprises 3 diagnoses involving sleep-related death in infants younger than 12 months, including sudden infant death syndrome (SIDS).[2] SUID is often used as a proxy measure for SIDS, indicating a diagnostic shift of some pathologists or medical examiners who may code a death as 1 of the other 2 diagnoses instead of SIDS, even in identical circumstances.[3–5]

The United States had the highest rate of SUID of wealthy countries in 2014 (0.87 per 1,000 live births).[6] Within the United States, SUID rates are extraordinarily high among non-Hispanic Black (NHB) infants (n = 1.85 per 1,000 live births) and non-Hispanic American Indian/Alaskan Native (AI/AN) infants (1.92), compared with non-Hispanic White (NHW) infants (0.82).[6] By contrast, SUID rates among Hispanic infants are much lower than among NHW infants (0.54), and US Asian infants have 1 of the lowest rates in the world (0.29).[6] Sudden infant death disproportionately affects low-income or oppressed communities in wealthy countries, and many of its most prominent risk factors overlap with circumstances associated with social disadvantage: low education, low income, tobacco use, alcohol use, young maternal age, unmarried status, inadequate prenatal care, as well as low breastfeeding rates,[7] with most potent risk factors including sofa sharing and bedsharing combined with tobacco or parental alcohol use.[8] Sudden infant death, including suffocation, is also more common among infants with low birth weight and prematurity,[9,10] which occur more commonly among communities with social disadvantage.

Breastfeeding is thought to be protective against SIDS,[11] likely through a combination of anti-infective, immunological, and nutritional properties of milk, maternal behavioral factors, and the physiologic response of sucking on infant arousal.[8] Compared with breastfeeding, infant formula feeding decreases infant arousability.[12] Videographic evidence shows that bedsharing, breastfed infants rarely sleep prone,[13–15] which decreases the risk of SIDS. In contrast to formula-fed infants, bedsharing, breastfed infants would rarely be placed near pillows that could suffocate them.[15] Bedsharing is associated with longer breastfeeding duration.[16,17]

The high rates of SUID among NHB and AI/AN infants and low rates among Asian Hispanic infants, relative to NHW infants, have been explained by combinations of risk factors.[7] AI/AN mothers have lower breastfeeding rates as well as relatively high rates of bedsharing, very high maternal antenatal smoking, preterm birth, and alcohol use.[7,18–22] Very low rates of breastfeeding, low rates of antenatal smoking, and very high rates of nonsupine sleep, bedsharing, and preterm birth were found among NHB mothers and their infants.[18–20,22] By contrast, relatively high breastfeeding initiation rates, low smoking rates, lower preterm birth rates, and higher rates of bedsharing, nonsupine sleep, and use of soft bedding for infants were found among Hispanic mothers and infants.[18–20,22] US Asian mothers have the highest breastfeeding rates, high rates of bedsharing and soft bedding for infants, and low rates of other risk factors.[3,18–20,22] A relatively high maternal antenatal smoking rate, a low preterm birth rate, and very low nonsupine sleep rate were found among NHW mothers and infants.[18–20,22]

One possible explanation for the difference in SUID rates between NHB and Hispanic infants may be breastfeeding; women from both groups have similar median incomes and labor characteristics, although far fewer families headed by single Hispanic mothers are in poverty (17% vs. 34%).[23,24] By examining the role of breastfeeding in the context of other known SUID risk factors among different US racial/ethnic groups, we can help guide policy-makers toward the most effective investments in reducing infant mortality in targeted racial/ethnic groups.

Overall, the explanation for the extremely high SUID rates among NHB and AI/AN infants remains poorly understood. To our knowledge, there have been no studies using individual-level data to explore the role of breastfeeding on SUID disparities. There has been some related research using population data[7] to show differences in breastfeeding and SUID rates by race/ethnicity. A computer-simulation study showed disparities in breastfeeding rates were associated with disparities in SIDS in NHB and Hispanic infants.[25] In 3 studies, researchers examined individual-level data to link breastfeeding to the decreased risk of SIDS[26,27]and infant death.[27,28] Birth certificate data were used in 1 study to show breastfeeding initiation was associated with lower rates of non-Black post-neonatal death in an urban population but not among a similar Black population.[28] However, a study of national birth certificate data showed breastfeeding initiation was associated with lower Black infant mortality rates among all age groups and was associated with lower SUID risk among all infants.[27]

In this study, we used 4 years of US birth certificate and linked death certificate data to examine the odds ratios (ORs) and absolute risk differences associated with never breastfeeding on SUID across US racial/ethnic groups. To examine our hypothesis that breastfeeding explains a substantial part of the racial/ethnic disparity in SUID, specifically between the 2 high-prevalence (AI/AN and NHB) and 2 low-prevalence (Asian and Hispanic) demographic groups, relative to the NHW group, we investigated mediation by breastfeeding in the association between race/ethnicity and SUID.