Sir.—Is there an acquired immunodeficiency syndrome (AIDS) embryopathy? The report by Marion et al1 raises two very significant questions on dysmorphology and the epidemiology of birth defects. What is sufficient evidence to establish that an agent causes a pattern of birth defects? How should potential confounding factors be controlled for when they are closely associated with exposure to the agent under study?
The features reported as characteristics of the AIDS embryopathy are mostly subjective. The evaluation of children for minor anomalies that are subjectively defined, such as prominent vermilion border, patulous lips, prominent forehead, and mild obliquity of the eyes, must take into consideration the possibility of observer bias. If the person examining the child is aware of the exposure to the hypothesized cause of the defect, he or she may unconsciously tend to scrutinize that patient more intensely than if the exposure status is unknown. In the study by