Hoffman weight for gestational age criteria, by gender and race


from: " Analysis of Birth Weight, Gestational Age, and Fetal Viability, U.S. Births, 1968" H. J. Hoffman, C.R. Stark, F. E. Lundin, and J.D. Ashbrook, Obstetrical and Gynecological Survey, Vol 29, No. 9, pp 651-681 , 1974
"Birthstat" is a SAS program to calculate newborn weight for gestational age status, by gender and race, using standard criteria based on a sample of U.S. births in 1968. (as published by Hoffman, et al (1974); as above). The program is offered by Karen Norberg, MD; e-mail norberg@nber.org

Issues to keep in mind:

  1. These criteria are based on a 50% sample of single, live births reported by 36 states and the district of Columbia in 1968, the first year of a revised US standard birth certificate that included report of date of last menstrual period.
  2. In spite of the large sample on which this table is based, (1,164,871 births), the 10th percentiles at extreme gestational ages are based on small numbers in this data; some nonlinearities probably reflect random variation. Other nonlinearities probably reflect true underlying effects in the population; for example, mean birthweights at very low gestational ages are higher than expected from a linear extrapolation from later gestational ages; this reflects the high mortality rate among very premature infants. Only those very premature infants of somewhat greater weight are likely to survive the birth process.
  3. These criteria represent a historical standard, against which later birth cohorts may be compared. Some "paradoxical" historical effects could occur. For example, improvements in medical care may result in lower rates of prematurity, and higher birth weight for a particular gestational age, but also higher rates of survival from prematurity, and higher rates of survival for lower weight infants at any particular gestational age. As a result, improvements in medical care could result, in theory, in higher rather than lower rates of prematurity and intrauterine growth retardation among live born infants.
  4. There is controversy about whether racial differences in birthweight in the U.S. are the result of economic or social factors, or reflect "true" biological variations among different ethnic populations. Investigators may wish to use the black/white distinction, or not, depending on the purposes of the study. However, differences in birthweight by gender are well established and clearly have a biological basis.
  5. There are other weight standards available; this table is chosen because it reflects a nationally representative sample, not selected for the presence or absence of prenatal illness or congenital malformations, and is the largest published sample. Other standards may be more appropriate for particular purposes .
This program calculates 10th and 25th weight percentiles by gestational age, gender and race, over gestational ages ranging from 22 - 52 weeks. Note: 1 gram = .035 oz; 1 oz = 28.349 grams
conventional cutoffs for gestational age are:
less than 37 or 38 weeks = premature;
more than 42 weeks = postmature;
Variable definitions:

sex 1 = male
2 = female

black 0 = other
1 = black

hs10 1 = equal to or below tenth percentile of weight for gestational age
0 = above tenth percentile

hs25 1 = equal to or below 25th percentile of weight for gestational age
0 = above 25th percentile
Download SAS code