Wrong Turns in Sudden Infant Death Syndrome Research

The 1995 conviction of Waneta Hoyt in Tioga County, New York, for the murders of her five children between 1965 and 1971 has resulted in considerable publicity.12 Much of the notoriety stems from the fact that in a landmark 1972 paper by Alfred Steinschneider published in Pediatrics these same babies were described as having succumbed to sudden infant death syndrome (SIDS).3 The appellation landmark is warranted because this paper has been quoted more than any other in the SIDS field—404 times between 1974 and 1996 in the journals surveyed by SciSearch4—and led directly to the widespread practice of using cardiorespiratory monitors to prevent SIDS. With justification the public is bewildered at how such a heinous crime could go undetected for so many years, and how a prestigious medical journal could publish a paper that had so much influence and yet turned out to be so wrong. The Hoyt case, along with others where homicide was thought to have been misdiagnosed as SIDS,5 has given new strength to those who have always suspected parents of being responsible for a high proportion of the deaths of infants who die suddenly and unexpectedly.Without intending to minimize the tragedy of innocent lives being taken, the question of undetected homicide should be viewed in the context of what was known about sudden unexpected infant death when these events occurred. The term “battered baby” was first put forth in 1962.6 It took at least a decade before the concept of infants being harmed by their care takers became generally accepted. Munchausen by proxy, of which the Hoyt case is a classic example, was not described until 1977,7 and even now the entity is insufficiently recognized by health professionals, let alone by judges and juries.Etched in my memory is an x-ray conference during my internship at Boston Children's Hospital in 1959 before the “discovery” of child abuse. A film was displayed of an infant with multiple fractures of varying ages. The debate among the eminent clinicians present was not about whether the diagnosis was a metabolic bone disease, but which one. The sole dissenter was the radiologist, Martin Wittenborg, who said simply, “This is trauma.” He was ignored. The classic paper by two distinguished pediatric neurosurgeons on subdural hematoma in infancy came from the same institution.8 Ninety-eight cases between 1937 and 1943 were described. With our present state of knowledge, most of those infants would be considered to have been abused, yet that possibility is never mentioned in the paper.Much is being made of the fact that Steinschneider's 1972 paper that spawned the apnea monitor movement was “authenticated” by the fact that it was published in Pediatrics. Nonsense! Only the most naive souls believe that medical journals have a lock on “the truth.” Angell and Kassirer recently said, “What medical journals publish is not received wisdom but rather working papers. Each of these is meant to communicate to other researchers and to doctors the results of one study. Each study becomes a piece of a puzzle that, when assembled, will help either to confirm or to refute a hypothesis.”9 Although I was critical of Steinschneider's advocacy for widespread use of apnea monitors,10 I felt then and still feel that his physiology studies attempting to identify infants at risk for SIDS were important to pursue.Steinschneider's is not the only article in the SIDS field to be wrong. In 1973 Naeye reported finding thickened pulmonary arteries in SIDS victims, which he suggested represented “tissue markers” of chronic hypoxemia prior to death.11 This paper, ranking second only to Steinschneider's in Scientific Search Citations (173),4 had a profound effect on SIDS research: Unfortunately it was equally flawed.12It is my subjective impression that over 80% of published papers about SIDS contain conclusions that have not been substantiated. The reasons are lodged in the twin scourges of SIDS research: imprecision in the diagnosis, and lack of appropriate controls.Much of the controversy about SIDS has centered on what it is, or even if it exists at all. Because the cause of SIDS remains unknown, and because there are no organisms, cell types, or biochemical markers that confirm the diagnosis, there are “lumpers” and “splitters.” Disagreement persists about the proportion of cases that fall into the category of known causes, and the proportion that fall into the unknown SIDS category.There is no disagreement that some infants who die suddenly and unexpectedly do so from known causes, and that one of those causes is homicide. The Hoyt cases demonstrate that it is possible to smother an infant where the postmortem findings are indistinguishable from SIDS. By means of covert video surveillance, Southall and colleagues have shown how it is done.13 Twenty years ago it was not unusual for parents whose infants died of SIDS to be accused of child abuse; it is now de rigueur for defense attorneys to invoke SIDS on behalf of those being tried for infanticide. Reese has outlined a helpful decision-making process for differentiating SIDS from abuse.14Speaking of being wrong, our Seattle group was among the first to identify the epidemiologic risk factors associated with SIDS.15 As the years went by, we were struck by the lack of risk factors in the vast majority of the SIDS victims we studied. This clinical impression led our pathologists and clinicians to reexamine and reclassify the 1065 cases previously diagnosed as SIDS, and compare these findings with birth certificate information. As suspected, we found that the social and/or demographic risk factors clustered in a relatively small proportion of infants, most of them with anatomic findings at autopsy, and that a significant proportion of classic-SIDS infants could not be distinguished from control infants.16 Two lessons emerged from this study: a) the need for precise and uniformly applied diagnostic criteria in the postmortem examination, and b) the need in SIDS research to concentrate on as homogeneous a population as possible, ie, the classic SIDS cases, rather than on the relative small proportion of “outliers.”The classic and most tragic example of “control failure” was the widespread practice in the early part of the 20th century of irradiating “large” thymus glands (called status lymphaticus) to prevent “crib death.” The thymus glands were felt to be enlarged because they were compared with the glands of infants whose thymus glands had involuted due to the prevalent causes of death of the time, infection and malnutrition. Appropriate controls, such as trauma victims, were few and far between. Lest we be too contemptuous of this misguided practice that produced an epidemic of thyroid cancer,17 we should recall that it was carried out by the most avant-garde physicians of the era.Most research in SIDS for the past 25 years has been carried out with three assumptions that have turned out to be false: a) the relationship of SIDS to apnea, b) anatomic evidence of preexisting hypoxia in SIDS victims, and c) trying to explain the relationship of epidemiologic risk factors to SIDS.My pessimistic view is that the mystery of SIDS will remain indefinitely because the research needed to produce answers is not being conducted. I doubt that there would be much argument among investigators in the field that SIDS is a developmental malfunction of centers that control respiratory and/or cardiac functions during sleep that occurs as the body's “thermostats” change from fetal to mature mode, usually between the second and third months of life. This thesis can only be proven or disproven by studying the physiology of tens of thousands of healthy infants in the requisite age group. Only in this way can variations of normal be identified with the hope of pinpointing infants at risk for SIDS. It is this type of prospective study of normal infants by Southall and his colleagues that discredited the apnea theory. They performed 24-hour tape recordings of electrocardiogram and breathing movements on approximately 10 000 London infants at the time of discharge from a maternity ward or neonatal intensive care unit and at home 6 weeks later. Twenty-nine of these infants subsequently died of SIDS. The SIDS victims did not show abnormally prolonged pauses in inspiration when compared with controls.1819Although the need for these types of large-scale prospective studies seems obvious, none are currently being conducted, and none are likely to be conducted in the future. Funding agencies are unwilling to make the necessary large financial investments; nor do capable investigators in need of academic advancement have the patience to tackle projects where results might not be forthcoming for many years. A multicenter collaborative study could produce results in a shorter time, but again, the leadership and money for such an endeavor are lacking.Americans tend to be uncomfortable with the unknown. We have a need for clear-cut answers. When these answers are not forthcoming, or turn out to be false, there is all too often an irresistible need to blame and find fault. It would be sad if publicity about the infanticide cases results in a return to the aura of suspicion that surrounded the families of SIDS victims in the past. A more appropriate memorial to the slain infants would be a nationwide effort to both strengthen our death investigation system for all infants who die suddenly and unexpectedly, and to increase support for research aimed at reducing the number of such deaths that occur.