An analysis of birth weight by gestational age using a computerized perinatal data base, 1975-1992.

OBJECTIVE

To develop birth weight-for-gestational age nomograms based on a computerized perinatal data base collected prospectively from 1975-1992.

METHODS

Using information from over 60,000 singleton deliveries (January 1975 through October 1992) at the MetroHealth Medical Center in Cleveland, Ohio, standard curves for normal birth weights were computed. Nomograms were developed for the overall population and for subgroups determined by factors known to affect fetal growth, including sex, race, smoking status, and gestational diabetes. The nomograms included the tenth, 50th, and 90th percentiles of birth weights for 24-44 weeks' gestation. Gestational age was based on clinical obstetric estimates confirmed by Dubowitz assessment of the neonate. In addition, third-order regression models were developed to predict median birth weight using gestational age. These models were validated using delivery data for the months of November and December, 1992, which were not included in model development.

RESULTS

The most significant predictors of median birth weight were the first-, second-, and third-order gestational ages, which explained over 80% of the total variation in birth weight. Other significant factors influencing birth weight included infant gender, maternal race, parity, smoking, and diabetes status. Among the marginally significant factors influencing birth weight were pay status and maternal age. In general, before 33 weeks' gestation, there were few differences in the birth weight percentiles of various groups except for those with diabetes; infants of diabetic women exhibited greater birth weights as early as 26 weeks' gestation.

CONCLUSIONS

Considering the large size of the data base and the diverse background of the study population, we believe that these nomograms provide useful norms of birth weight for an indigent urban population. These norms enhance the obstetrician's and neonatologist's ability to identify true cases of retardation or acceleration of intrauterine growth. Simple mathematical models provide easy calculation of the median birth weights for 24-44 weeks while adjusting for many confounding factors.