1 Early Weaning and Perinatal Smoking Findings from

1 Early Weaning and Perinatal Smoking Findings from the 2000 -01 Oregon PRAMS Jihong Liu, Sc. Da, b, Kenneth D. Rosenberg, MD, MPHb a ORISE, b Division of Reproductive Health, CDC Office of Family Health, Oregon Department of Human Services The 9 th Annual Maternal and Child Health Epidemiology Conference Tempe, Arizona, December 10 -12, 2003

2 Literature review (1) è Compared to non-smokers, women who smoke are: Less likely to initiate breastfeeding Likely to breastfeed for a shorter duration P. D. Hill and J. C. Aldag (1996), L. H. Amir and S. M. Donath (2002)

3 Literature review (2) è Possible mechanisms to explain association between maternal smoking and decreased breastfeeding duration: Nicotine reduces the level of prolactin Smoking interferes with the release of oxytocin Milk volume of smokers is significantly less Breast milk of smokers has lower fat content Infants of smoking mothers have more problems with breastfeeding A. N. Anderson et al. (1982), A. N. Anderson et al. (1984), F. Vio et al. (1991), J. M. Hopkinson et al. (1992), L. H. Amir (2001)

4 What is missing in the literature? è è Few studies address changes in smoking habits around pregnancy and the synergistic effect with the intensity of postpartum smoking Few studies use population-based samples

5 Objectives è è To examine whether postpartum cigarette smoking has a stronger impact on breastfeeding duration than prenatal smoking To study whethere is a dose-response effect between smoking status and breastfeeding duration

6 Data è Data from the 2000 -01 Oregon PRAMS were used Stratified random sample of recent mothers from birth registry Response rate: 72. 6% (n=3, 215) For more information, visit URL http: //www. ohd. hr. state. or. us/pch/prams/index. cfm

7 Methods (1) è Calculate percentage of mothers who are not breastfeeding at 10 weeks All mothers were interviewed when their babies were 10 weeks old or older Weighted for oversampling, nonresponse and noncoverage è Kaplan-Meir life table method Probabilities of weaning by postpartum week Censoring at interview time

8 Methods (2) è Logistic regression models Outcome § Not breastfeeding at 10 weeks Predictor § § § Self-reported smoking status during three trimesters of pregnancy and the postpartum period Intensity of postpartum smoking (< or 10 cig/day) Six mutually exclusive categories n n n Never smoked Quitters (who stayed quit) Postpartum relapsers (< 10 cig/day) Postpartum relapsers ( 10 cig/day) Persistent smokers (< 10 cig/day) Persistent smokers ( 10 cig/day)

9 Methods (3) è Logistic regression models Control variables § § § è Mother’s age, education, marital status, race/ethnicity, parity, insurance status at delivery Child’s sex and birth weight Pregnancy intention Use of prenatal care at the 1 st trimester Self-reported barriers for breastfeeding (taking medicine, plan to go to work or school) SAS 8. 0 and SUDAAN 8. 0

10 Percentages of not breastfeeding at 10 weeks by smoking status

11 Probabilities of weaning by postpartum week

12 Odds ratios of weaning by 10 weeks by perinatal smoking status

13 Limitations è è è Retrospective report of smoking status without biochemical verification No information on the timing of smoking relapse during postpartum period No information on employment status

14 Conclusions è è Heavy smoking (10 or more cig/day) during the postpartum period tends to be associated with early weaning Non-smokers, quitters, and postpartum relapsers who smoked less than 10 cig/day may have similar odds of early weaning

15 Public health implications è è è Provide information on the adverse impact of postpartum smoking on breastfeeding Assist pregnant women who wish to stop smoking during postpartum period Smoking relapse prevention may increase the proportion of women who breastfeed their babies longer

16 Acknowledgements Alfredo P. Sandoval James A. Gaudino Patricia R. Westling Tina M. Kent Laura A. Zukowski