Fifty years ago, an extraordinary group of pathologists, pediatricians, and epidemiologists met to address the limited research and numerous diagnoses applied to unexplained, largely sleep-related pediatric deaths. Their efforts led to the identification of a new category of disease, sudden infant death syndrome (SIDS). Among the observations fueling the creation of the category was the relative sparing of neonates despite what should be increased vulnerability in the early days of life. When Dr Bruce Beckwith, the conference’s leader, asked whether anyone doubted this age-related sparing, not a single objection was raised.1 Yet, humility about what was not yet known led to broader diagnostic criteria of “any infant or young child.” Years later, SIDS was redefined as pertaining to infants <1 year of age.2 Still later, when classification practices scattered because of objections about the concept of SIDS and the varied use of alternative terminologies, the composite of sudden unexplained infant death (SUID) came to stand in.3
In this issue of Pediatrics, big data concur with the pioneering thinkers and draw the demarcation at 1 week. In “Distinct Populations of Sudden Unexpected Infant Death Based on Age,” Lavista Ferres et al4 used elegant, unbiased computational modeling without predetermined age boundaries to determine distinct age-related subgroups in SUID. Using profiles that included birth order, birth weight, marital status, maternal age, and smoking, they showed that associated features of SUID during the first week of life were statistically divergent from SUID during the remainder of the first year. Their efforts join other recent contributions5,6 arguing that this earliest unexplained neonatal mortality may represent something different from SIDS and its analogs. At first glance, this accumulated research is reminiscent of findings that led to the demise of what was once called “near SIDS,” when differences in risk factors and incidence over time led to the conclusion that apparent life-threatening events, now called brief resolved unexplained events, were unrelated to SIDS.7
This analysis centers on birth and death statistics in early life and is especially reliant on day 0 data. But accounting for mortality near the time of birth is notorious for its challenges. The definition of live birth is broad, and accounting for intrapartum and neonatal death is tangled.8 The World Health Organization’s definition of a live birth requires breathing or “any other evidence of life, such as beating of the heart, pulsation of the umbilical cord, or definite movement of voluntary muscles, whether or not the umbilical cord has been cut or the placenta is attached,”9 which is a permissive category. Half of all stillbirths occur during labor and often involve highly emotional intrapartum complications and unpredictable odds for survival. They involve delivered newborns who cannot be resuscitated and resuscitated newborns who live only briefly. More than 60% of deaths in the first month occur on the first day, largely in the first hour.10 This leads to inconsistencies as to which deceased infants receive birth certificates (and are thus included in mortality statistics) and which do not. In Lavista Ferres et al’s4 research, 47% of the infants in their distinct early mortality population died on day 0. Compared with the human drama in these moments, clean data are a small consideration. Nonetheless, the quality of the data makes conclusions derived from them hazardous.
The past 50 years of research have led to other advances that are hard to reconcile with their conclusions. Research on SIDS and stillbirths suggests a biological continuum between fetal vulnerabilities and SIDS, not a disconnect.11,12 Although the researchers showed that controlling for the effects of prematurity and low birth weight did not affect their analysis, there is a substantial body of research showing that pregnancies before and after an infant who experienced SIDS have higher rates of poor fetal growth and early labor and that risk for SIDS can be calculated from obstetric characteristics.13,14 Elevated α fetoprotein levels during pregnancy are associated with both stillbirth and SIDS.15 Stillbirth and SIDS share certain brain abnormalities16 and variants in cardiac arrhythmia genes.17 Although Lavista Ferres et al’s4 research may point to an area that requires new thinking, misclassification may exaggerate the differences within SUID and mask these commonalities. In fact, their study’s identification of birth order differences in the 2 populations may reflect the association between SIDS and less successful later pregnancies.
In the flawed and contradictory universe of perinatal data, the best way forward is a continued struggle for better definitions, careful analysis, and the integration of varied research findings. This requires serious inquiry and a healthy dialogue, without overstatement. To whatever degree distinctions are unduly influenced by weaknesses in current methods of surveillance, the exchange of different ideas and conclusions sharpens our questions, and that is hard-earned progress on a difficult problem. In “Distinct Populations of Sudden Unexpected Infant Death Based on Age,” Lavista Ferres et al4 have made an important contribution to the conversation.
Opinions expressed in these commentaries are those of the author and not necessarily those of the American Academy of Pediatrics or its Committees.
FUNDING: No external funding.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2019-1637.
POTENTIAL CONFLICT OF INTEREST: The author has indicated he has no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The author has indicated he has no financial relationships relevant to this article to disclose.