, 1997). Demonstrating associations with changes taking place between pregnancies permits stronger causal inferences. Also, applying the longitudinal study design to examine the influence of increased parity for subsequent smoking behavior in a substantial sample of primiparous women is a significant strength. Moreover, www.selleckchem.com/products/brefeldin-a.html in addition to adjusting for the most important risk factors for maternal smoking, the possibility to take into account prior smoking behavior when investigating risk factors for smoking in a later pregnancy strengthens the findings. Because both women who quit and women who do not quit smoking share a history of smoking, their situation may deviate substantially from that of nonsmokers (Wakschlag et al., 2003). As postpartum relapse rates are substantial (Colman & Joyce, 2003; Hajek et al.

, 2001; Kahn et al., 2002), many pregnancy quitters may have suspended smoking rather than actually quit, and as such, their situation may influence the associations being investigated. Still, some limitations need to be considered. One possible limitation of the study is that smoking was assessed by self-report. Social desirability and the stigma attached to smoking during pregnancy may result in biased estimates, and validation of self-report by cotinine measurements typically demonstrates an under-reporting of actual smoking during pregnancy (Dietz et al., 2011; Pickett, Rathouz, Kasza, Wakschlag, & Wright, 2005).

However, in a substudy of the MoBa cohort, it was found that self-reported smoking status during pregnancy had a sensitivity of 82% and a specificity of 99% in contrasted to plasma cotinine concentrations, indicating that self-reported smoking is a valid marker for smoking behavior in the MoBa cohort (Kvalvik et al., 2012). Another limitation of the study sample is selection bias from the larger population from which the women were sampled and a selected sample of women participating with two pregnancies. A comparison of women taking part in MoBa with all women giving birth in Norway identified under-representations such as women younger than 25 years and single women (Nilsen et al., 2009). However, although differences in prevalence estimates were observed, no relative differences in relevant exposure-outcome associations were identified. Nevertheless, the associations obtained in this study should be replicated in more representative and culturally diverse samples.

Whereas many previous cross-sectional studies have identified important yet relatively stable demographic and contextual risk factors for maternal smoking, this study offers an important distinction in identifying also more modifiable factors associated with smoking behavior prior to and during pregnancy. Pregnancy and early parenthood are critical periods Carfilzomib during which substantial social changes take place, changes that involve both the woman herself and her partner (Hildingsson, Tingvall, & Rubertsson, 2008).