Introduction

In Chapter 1 the various definitions of sudden infant death syndrome (SIDS) were discussed, with the one common theme being the lack of diagnostic features. In a way, pathology represents the weak link in the SIDS chain, as there have never been consistent and reproducible diagnostic tissue markers (1, 2). Thus, current definitions of SIDS are generally of exclusion, which means that the term “SIDS” can only be used for an infant death once other causes of sudden death have been excluded. This requires very careful interpretation of the autopsy findings to determine whether tissue changes are causative or coincidental, or whether they are merely epiphenomena. Other problems also involve the history and circumstances, which may be typical but are also not diagnostic (3, 4).

As there are no pathognomonic markers at autopsy in SIDS deaths, there is a danger that the “diagnosis” will be used inappropriately for natural deaths, accidents, and homicides — John Emery's “diagnostic dustbin” (5, 6). The degree of certainly with which alternative diagnoses can be made varies and the Avon system grades this from Ia for completely unexplained deaths, to grade III when a definite cause can be established (7).

Another issue which has arisen many times is the lack of a control population with which to compare the findings in SIDS autopsies. Examples of this include the reported changes in the heart in SIDS infants which were subsequently considered to be part of normal growth and development of an immature organ, and the ubiquitous finding of minor inflammatory infiltrates in various organs that are often more common in controls and do not, therefore, indicate significant lethal occult infection (8, 9).

Despite recommendations for decades that autopsies are a mandatory part of the work-up of SIDS cases, they have not always been performed in cases that have still been classified as “SIDS”. For example, in publications from the 1990s the autopsy rate was less than 25% in Belgium (10) and autopsies occurred in only 50 to 60% of infants in the Netherlands (11).