Birth weight traditionally is classified using population-based, sex-adjusted centiles. An infant weighing less than the 10th percentile at birth is considered to be small for gestational age (SGA). Some of these infants, however, are constitutionally small but not growth-restricted. This study was an attempt to reclassify infants using customized and population percentiles, and to compare perinatal morbidity rates in those classified as SGA and those appropriate for gestational age. The cohort study included 374 SGA pregnancies from a general obstetric population of 12,879. Pregnancy outcomes were compared for “SGA both” infants who were below the 10th percentile by both population and customized percentiles; “SGA customized only”; and “SGA population only.” All SGA pregnancies were evaluated by an ultrasound at less than 20 weeks gestation and by umbilical Doppler study within 2 weeks of delivery. The National Women’s Hospital database of births in the years 1993–2000 was used to study the use of customized percentiles in a general obstetric population. Customized SGA was the customized birth weight below 10th percentile using the New Zealand percentile calculator. Population SGA was sex-adjusted birth weight less than 10th percentile and non-SGA was birth weight of 10th percentile or more by both customized and population parameters.
At delivery, 80% of infants were SGA by customized percentiles and 81% were SGA by population percentiles. The proportion of infants who were SGA by both these percentiles was 72%. In addition, 7% of infants were “customized SGA only,” 9% were “population SGA only,” and 12% were “non-SGA both.” Mothers of infants who were reclassified as “population SGA only” had lower body mass indices than those who were “SGA both,” and they also were likelier to be nulliparous. Women in the “customized SGA only” group tended to have higher body mass indices. Rates of preeclampsia, abnormal Doppler findings, and cesarean section for fetal distress all were greater in the “SGA both” and “customized SGA only” groups than in the “non-SGA both” group. These forms of morbidity were similarly frequent in the “SGA both” and “customized SGA only” groups. Indices of newborn morbidity, including a long hospital stay and composite morbidity, were increased to similar degrees in the “SGA both” and “customized SGA only” groups. Morbidity was comparatively low in the “population SGA only” and “non-SGA both” groups. All seven perinatal deaths were in the “SGA both” group. Morbidity risk in babies who were “customized SGA only”—and who would not have been detected using population percentiles—had high morbidity rates comparable to those of infants in the “SGA both” group. Compared with the “non-SGA both” group, relative risk figures for perinatal death were high in the “SGA both” and “customized SGA only” groups.
It appears that the use of customized birth weight percentiles is able to identify small infants who are at increased risk of perinatal morbidity and mortality.