Bias in Studies of Preterm and Postterm Delivery Due to... : Epidemiology (original) (raw)
ORIGINAL ARTICLES: PDF Only
Bias in Studies of Preterm and Postterm Delivery Due to Ultrasound Assessment of Gestational Age
Henrihsen, Tine Brink12; Wilcox, Allen J.3; Hedegaard, Morten2; Jørgen Secher, Niels2
From the 1Danish Epidemiology Science Centre, Institute of Epidemiology and Social Medicine, Aarhus University, Aarhus, Denmark; 2Perinatal Epidemiolog-ical Research Unit, Department of Obstetrics and Gynaecology, Aarhus University Hospital, Aarhus, Denmark; and 3Epidemiology Branch, National Institute of Environmental Health Sciences, Research Triangle Park, NC.
Abstract
Ultrasound measurement of fetal dimensions is widely used for estimating gestational age. A little-discussed limitation of this method is that variations in fetal size at a given stage of pregnancy are converted to differences in gestational age. Factors that affect pregnancy duration often affect fetal size also. We explore how the effect of such factors may he biased when gestational age is determined by ultrasound. We selected 3,606 women with singleton pregnancies (1989–1991) who had an early ultrasound measurement of fetal biparietal diameter (BPD) and a good-quality history of last menstrual period (LMP). Using the two measures of gestational age, we estimated risk of preterm and postterm delivery for female babies vs males and for smoking women vs nonsmoking women. There was a 13% excess of preterm delivery among female babies when gestational age was determined by ultrasound, but no excess when gestational age was defined by LMP. For postterm delivery, female babies had a 19% lower risk with ultrasound-defined age, but no deficit with LMP-defined age. We found a similar bias with ultrasound in the analysis of maternal smoking. Thus, factors that reduce fetal size inflate the risk of preterm delivery and deflate the risk of postterm delivery when gestational age is based on ultrasound measurement of the fetus. This bias can distort the relative risk of preterm or postterm delivery by 10–20%. (Epidemiology 1995; 6:533–537)
© Lippincott-Raven Publishers.