Errors in gestational age: evidence of bleeding early in pregnancy.

OBJECTIVES: This study explored the extent of errors in gestational age as ascertained by last menstrual period. METHODS: More than 1.5 million birth records (covering the years 1967-1994) from the population-based Medical Birth Registry of Norway were used to study variation in gestational age within strata of birthweight. RESULTS: Within 100-g strata of birthweight, it was found that the observed gestational age distribution could be divided into 3 distinct underlying distributions separated by approximately 4 weeks. This pattern was present through all birthweight strata, from 200 g up to 4700 g. In addition, the apparent misclassification causing a gestational age 4 weeks too short was much more common among low-birthweight births than among heavier births. CONCLUSIONS: The separation of the gestational age distributions by intervals of close to 4 weeks suggests that errors in gestational age measurements are caused by factors related to menstrual bleeding. Furthermore, there is evidence for a strong relation between bleeding at the time of the next menstrual period after conception and low birthweight. This conclusion should be approached with caution because of the retrospective nature of the data.