A Commentary on Maternal Smoking During Pregnancy: The Story Behind the Trends

The adverse effects of maternal tobacco smoking during pregnancy on maternal and infant health have been studied for decades, ¹ and the prevalence of maternal smoking overall and during pregnancy has declined considerably since publication of the landmark Surgeon General’s report in 1964. ² Smoking is associated with an array of adverse pregnancy outcomes, including low birth weight, preterm birth, small for gestational age, stillbirth and infant mortality, and several birth defects. ³ However, many women continue to smoke before and during pregnancy, despite long-running public health education campaigns and health provider recommendations during prenatal care to enter smoking cessation programs. One of the goals of Healthy People 2020 is to increase rates of abstinence from cigarette smoking among pregnant women, with a target of 98.6%. ⁴ Although this goal does not specify whether maternal smoking status should be measured at delivery or at some other time during the pregnancy, the message that maternal smoking poses a risk to the infant could not be clearer.

Smoking during pregnancy and smoking among women of reproductive age can be measured through several national sources of public health data, including the natality data collected through reporting on the US Standard Certificate of Live Birth, ⁵ the Pregnancy Risk Assessment Monitoring System (PRAMS) (https://www.cdc.gov/prams), the National Survey of Family Growth (NSFG) (https://www.cdc.gov/nchs/nsfg/index.htm), and the Behavioral Risk Factor Surveillance System (https://www.cdc.gov/brfss/index.html). All these sources rely on self-report by mothers or women of reproductive age. The study of temporal trends in maternal smoking poses challenges for these data sources. For example, data on maternal smoking during pregnancy on birth certificates was initiated in 1989, but questions related to smoking reduction or cessation were introduced only with the 2003 revised birth certificate. Ananth et al ⁶ conducted an age–period–cohort analysis using national birth certificate data to compare changes in maternal smoking patterns by age between white and black women. PRAMS started in the late 1980s in a few states and has gradually added states to achieve almost national coverage. NSFG also provides data on trends in cigarette smoking during pregnancy but only for the past 20 years. The estimated prevalence of smoking was 13.4% during 1997-2002, 13.9% during 2006-2010, and 12.3% during 2011-2015. ⁷ The NSFG data show a smaller decline in smoking prevalence than found in other data sources. Thus, none of these sources provides data sufficient to analyze national trends in maternal smoking or smoking cessation during the past 3 decades.

This issue of Public Health Reports includes an article by Hansen et al ⁸ on changes in the prevalence of maternal smoking and smoking cessation during pregnancy from 1985 to 2014. The authors used data from the Early Childhood modules of the National Health and Nutrition Examination Survey (NHANES) in 1999-2014; these modules provide data, obtained from interviews, on children from birth through 15 years of age, including data on the smoking habits of the mother while pregnant with the child. Mothers usually served as proxies during the interviews. Although this source of data is potentially limited by recall bias, ^9,10 it does provide consistently measured data on maternal smoking behavior from a representative sample that provides national estimates.

Although the annual estimates of the maternal smoking rate and the maternal smoking quit rate reported by Hansen et al ⁸ have wide 95% confidence intervals (CIs), the study showed substantial declines in the prevalence of maternal smoking, from 25.7% (95% CI, 15.3%-36.0%) in 1985 to 10.1% (95% CI, 7.1%-13.0%) in 2014. The proportion of women who previously smoked but quit smoking at any time during pregnancy increased from 36.6% (95% CI, 20.3%-52.9%) to 54.9% (95% CI, 44.4%-65.4%) during the same period. A study of national birth certificate data in 2014 by Curtin and Mathews ¹¹ reported that 8.4% of women who gave birth smoked at any time during their pregnancy; this percentage is within the 95% CIs reported by Hansen et al. However, Curtin and Mathews reported a rate of smoking cessation during pregnancy of 20.6% in 2014, which was near the lower limit of the CI shown for 2008 in Figure 1 of Hansen et al. ⁸ The differences in these estimates may reflect differences in how questions were asked, how quit rates were calculated, or recall bias.

It would be interesting to learn how the NHANES data might be used to examine associations between maternal smoking behavior, including smoking cessation during pregnancy, and child health status and outcomes. As Hansen et al ⁸ demonstrate, NHANES can provide estimates of family income, which, when analyzed in concert with maternal race/ethnicity and age, could provide useful data for targeted interventions. Hansen et al do not fully explore patterns of smoking behavior within racial/ethnic groups. Ananth et al, ⁶ using data from 1990 through 1999, a period that overlaps with the period studied by Hansen et al, showed that among age groups of white mothers, adolescent mothers were the most likely to smoke, and as age increased, smoking rates declined; whereas among age groups of black mothers, the prevalence of smoking during pregnancy increased with maternal age. Race-specific analyses might show important differences in secular trends.

On the other hand, researchers should not discount the versatility of data obtained from birth certificates, which now support analysis of patterns by trimester and can identify incident smoking (ie, women who initiate smoking during pregnancy). Vital records can be linked to other administrative databases, including systems that collect data on hospital discharges, Medicaid, newborn screening, and birth defects, and cancer registries.

Hansen et al ⁸ provide data on national secular trends in maternal smoking and smoking cessation during pregnancy during the past 30 years, demonstrating both public health progress and opportunities for improvement. However, there is more work to be done.