Background: Most birth weight-for-gestation charts are based on hospital populations or on birth certificates. In Flanders, the northem region of Belgium, we have aimed at obtaining standards of birth weight-for-gestation of the total population. The data were collected by the Study Center for Perinatal Epidemiology, that registers 98% of all deliveries.

Subjects: For each hospital delivery in Flanders: gynecologists, midwifes, and pediatricians register 30 items including birth weight and gestational age. We studied the records of all registered single, life-born infants with certain post-menstrual age, free of malformations and born between 1987 to 1995. Percentiles of birth weight-for-gestation were constructed from 24 to 42 weeks of gestation.

Results: The study population consists of 429.070 infants, after exclusion of 113 infants with apparent registration errors. Compared with the Denver standards of birth weight-for-gestation, percentiles were shifted to lower values at early gestational ages, but reached much higher weights at term (p10 = 2960 at 40 w as compared to 2630 g for the Denver standards).

Conclusion: Birth weight-for-gestation standards should rely on recent and preferably population-data. The present standards fit well with those of the UK Northem Region (Tin el al. Brit J Obstet Gynecol, 1997, 104, 180-185), based on a comparable population and a similar registration protocol.