Abstract
Definition by exclusion is sometimes unavoidable but never really satisfactory. It is particularly unsatisfactory in the case of SIDS because some specific causes of unexpected infant death, such as suffocation or metabolic disorder, may leave little or no obvious sign. There are those who would argue that SIDS is not a proper diagnosis anyway but just a cloak for our diagnostic penury: there is always a specific cause of death if only we knew how to find it. In addition there remains the problem of what degree of necropsy findings constitutes an adequate cause of death. In the SUDI study there were several instances in which the expert regional panel, guided by their paediatric pathologist who had seen the slides, disagreed with the interpretation of the reporting pathologist, upon whose opinion the cause of death had been registered. Usually the reporting pathologist thought his findings constituted an adequate cause of death while the panel did not; this occurred in at least one in 10 of unexpected deaths not registered as SIDS (the exact proportion is uncertain because panels did not always record this point). Consistent errors of this sort will falsely lower the SIDS rates recorded by the Office of Population Censuses and Surveys (OPCS, now the Office for National Statistics), the distortion being greater now that deaths are fewer. The subjectivity inherent in the definition of SIDS means that some disagreement is inevitable, but it could be kept to a minimum if necropsies on infant dying unexpectedly were done only by specialist pathologists. Meanwhile all pathologists who do infant necropsies should at least try and achieve better consensus.