Divergent trends in maternal cigarette smoking during pregnancy: United States 1990-99.

While examination of causes for trends in smoking have largely focused on how changes have occurred with maternal age and, less commonly, time period, little is known as to how age, period and birth cohort interact on trends in the prevalence of smoking during pregnancy. We performed a population-based, retrospective cohort study based on the vital statistics records comprised of White (n=24,499,629) and Black (n = 5,096,625) women delivering in the United States in 1990-99. Smoking prevalence rates were derived by seven 5-year maternal age groups (15-19 to 45-49 years), two time periods (1990-94 and 1995-99), and eight 5-year maternal birth cohorts (1945-49 to 1980-84) after adjusting for the confounding effects of gravidity, education, marital status, and lack of prenatal care through multivariable logistic regression models. The prevalence of smoking was 17.3% among Whites, and 13.5% among Blacks, with substantial variations by age, time period, and birth cohort. The rate declined with increasing age among Whites during the 1990-94 and 1995-99 periods. Among Blacks, the rates increased steeply with advancing age up to 25-29 years and began to decline thereafter. Smoking rates declined among both Whites and Blacks with increasing birth cohort within each age strata. These rates were highest among multigravid women (gravida > or = 2), and lowest among primigravid women. The rates among Whites declined with increasing maternal age for each gravida. Among Blacks, smoking rates for each gravida increased with advancing age up to 25-29 years, and plateaued among older women. Variation in smoking prevalence by age, time period, and birth cohort provides impetus for designing interventions to reduce smoking. Such studies should not only consider cross-sectional trends, but also the divergent patterns by age and cohort among women of different race/ethnic groups and gravidity.