| Hypoglycemia is a serious problem experienced by many newborn infants during the early postbirth period, and if left untreated can cause serious long-term adverse neurologic sequelae such as learning disabilities and mental retardation. Because of the seriousness of hypoglycemia, early identification and appropriate treatment during the early postbirth period is important. Historically, size at birth has been an important factor in determining which infants are screened for hypoglycemia during the early postbirth period. However, the use of anthropometric measurements, singly, in combination, and in anthropometric indices and ratios, to predict hypoglycemia has not been studied adequately to determine if size at birth is a valid predictor of an infant's risk of hypoglycemia during the early postbirth period. Anthropometric measurements of weight, head circumference, chest circumference, abdominal circumference, mid-arm circumference, thigh circumference, and length measurements were obtained on a convenience sample of 94 Caucasian and 63 African-American infants within 24 hours after birth. Glucose measurements were obtained at two hours of age.; Two way ANOVAs demonstrated that there were significant differences by race in weight, head circumference, chest circumference, abdominal circumference, mid-arm circumference, and length measurements, and by gender in thigh circumference measurements. Logistic regression statistics were used to calculated odds ratios for single anthropometric measurements, anthropometric indices and ratios, models of multiple anthropometric measurements, and measurement-by-gestational age scales. The findings demonstrated that weight measurements [odds ratio 1.04 (0.778--1.398)] were not predictive of risk of hypoglycemia, and that the head circumference measurements has the highest odds ratio [1.42 (0.959--2.130)]. The mid-arm circumference/head circumference ratio was one of the best predictors [.45 (0.244--0.834)], because the proportion of infants with a ratio between 0.27--0.29 was much higher than at other points. Also the model containing both head circumference measurements and mid-arm circumference measurements was a much better predictor than either of the two measurements singly. The addition of the following variables race, gender, race x gender, and gestational age did not add to any of the models. |
