Impact of race and tobacco use on birthweight by gestation week

We use innovative statistical methods to evaluate the impact of maternal race and tobacco use on birthweight (BWT) across gestation week (GW). Using NC Detailed Birth Records for 1999-2003, we analyze 206,379 singleton non-anomalous first births born to women aged 15-44 with no alcohol use or medical complications, limited to non-Hispanic white (NHW), non-Hispanic black (NHB), or Hispanic (H) race. Separate linear models were fit for GW 26-42, estimating BWT with maternal tobacco use, age, education, marital status, and race and infant sex as explanatory variables. Separate collections of models were fit for racial subgroups.

Compared to NHW, NHB race is associated with statistically significant lower BWT (~96g-181g) starting in GW 31, while H ethnicity has statistically significant lower BWT (~67g-99g) starting only in GW 37. The absolute decrement in BWT increases with GW, but is relatively constant in proportional terms: ~5% (~2.5%) of mean BWT for NHB (H). Tobacco use is associated with statistically significant lower BWT starting in GW 31 for the entire study population and NHW subgroup, but not until GW 36 for the NHB subgroup. Again, the absolute impact of tobacco use grows from 104g in GW 32 to 176g in GW 40, but the proportional impact stays relatively constant at ~5% of total BWT. The proportional impact of tobacco use is ~5.5% (~3%) for NHW (NHB).

Third-trimester fetal surveillance may be indicated for at-risk populations because some factors associated with reduced infant BWT do not manifest their effect until relatively late in gestation.