Possible diff erences in fetal size by racial origin

In The Lancet Diabetes & Endocrinology, Jose Villar and colleagues describe an international study of fetal ultrasound biometry (widely used to assess fetal growth) through pregnancy, aimed at producing a global standard for fetal growth. The investigators enrolled 4607 pregnant women who were educated, affl uent, and healthy—to minimise variation in fetal growth attributable to diff erences in socioeconomic and environmental status—in cities in Brazil, Italy, Oman, UK, USA, China, India, and Kenya. The researchers measured fetal size carefully through pregnancy, and concluded that the population characterised could be used to produce an international growth standard, which could be used worldwide. But were their methods appropriate and their conclusion justifi ed? Although Homo sapiens is a single species, medically important genetic diff erences have developed during the evolution of geographically separated populations. For example, the diff erent racial distributions of sickle cell anaemia and cystic fi brosis are widely appreciated. Major variations in female size also arise; in Africa reported average female heights vary from as high as 1·8 m (some Sudanese and Maasai tribes) to as low as 1·33 m (forest people [pygmies]). Genetic diversity is much less pronounced outside Africa, which is attributed to the population bottleneck—a small cohort of our species left Africa and subsequently expanded in numbers—but substantial size variation still exists in other areas. For example, the average female height is 1·67 m in Sweden and only 1·51 m in Bangladesh. A particular concern in this study is the exclusion of women with a maternal height less than 153 cm, which the authors say excluded 13% of otherwise eligible participants, mostly in India and Oman. Therefore, the studied participants were not representative of the average genetic background in these two sites. Although most human genetic variation arises in Africa, all but one of the study sites were outside Africa, which restricted the genetic diversity that could be recorded. Birthweight is strongly linked to maternal size, and therefore it is not surprising that birthweight also varies between racial groups, even allowing for racial diff erences in gestational length. Therefore, the stage of pregnancy at which racial diff erences in size are detectable on ultrasound assessment is important to establish, so that norms appropriate for each racial group can be determined. Serial ultrasound fetal biometry from early pregnancy onwards allows detection of true fetal growth restriction (a decrease in the centile size of an individual fetus, as compared with the population average, as gestation advances). However, if early scans are not available, the measured size of a fetus should be compared with an appropriate norm. This situation has led to the development of customised centiles, which take racial group into account. Villar and colleagues point out that it is important to distinguish between genetic (racial, nonmodifi able) and environmental (socioeconomic, modifi able) eff ects on fetal growth. Ideally then, their study of ultrasound biometry through pregnancy in eight study sites would have compared the eff ect of variations in environmental factors (eg, nutrition, smoking, and work activities) in clearly defi ned racial groups. However, the authors decided on a diff erent approach; to study only educated, affl uent, healthy women aged between 19 and 34 years, and taller than 153 cm. Because their study sites were mainly cities, their affl uent population is likely to have been racially heterogeneous. Therefore, perhaps unsurprisingly, the investigators noted little diff erence in fetal ultrasound biometry growth profi les between the sites, although fetuses in Italy had a consistent roughly 0·7 SD bigger head circumference than did fetuses in China from 14 weeks gestation onwards. Researchers concluded that diff erences in fetal growth and newborn size are more likely due to environmental and socioeconomic diff erences than genetic variation. However, the authors also note that study patients did not undergo genetic profi ling, self-identifi ed racial origin is not reported, and recruitment was restricted to educated, affl uent, and well-nourished women. Together, these limitations mean that the results of this study could not be used identify any major diff erences in response to adverse environmental factors between racial groups. For example, advanced maternal age and obesity are more likely to increase gestational diabetes and birthweight in south Asians than in white Europeans. Villar and colleagues’ conclusion does not exclude important racial diff erences in functional Published Online July 7, 2014 http://dx.doi.org/10.1016/ S2213-8587(14)70157-3