Perinatal mortality by birth order within cohorts based on sibship size.

Cross-sectional surveys of perinatal mortality show a U-shaped curve when plotted against parity, implying that fourth and subsequent babies are at increased risk. Our study of a large, population-based longitudinal data set shows that this result is an artefact and that perinatal mortality falls with increasing parity. Within cohorts of mothers based on attained sibship size the perinatal mortality decreases with increasing parity and increases with sibship size. These associations, which are not noticeably affected by maternal age, ssem in part to operate through an association between parity, sibship size, and birth weight. This analysis shows the importance of using longitudinal data in analysing such relations.