This research reveals several important findings. First, there are differences in the relationship between ACEs, psychological distress, and adult smoking by sex. While the relationships between ACEs and psychological distress was evident among both men and women, after adjusting for covariates, there were not significant relationships between ACEs and smoking or psychological distress and smoking (after each ACE was added to the model) among men. Second, this research suggests that women, particularly those who have experienced emotional or physical abuse, physical neglect, or parental separation or divorce as children may be at particular risk for smoking in adulthood. In fact, approximately 22% of the relationship between emotional abuse and adult smoking was mediated though psychological distress as was 17% of the relationship between physical abuse and adults smoking, 14% of the relationship between physical neglect and adult smoking, and 10% of the relationship between parental separation and divorce and adult smoking.
Our findings confirm results of earlier research suggesting sex differences in smoking behavior and pattern . Although negative affect, including depression, is related to smoking among both men and women, the relationship is much stronger for women [88–90]. In fact, recent research suggests that stressful childhood life events may disproportionately influence a women’s decision to use drugs [48–57]. This may be due in part to differences in coping styles and socialization ; females may develop more passive styles of responding to threats and distressing events as opposed to boys who may engage in a more active coping style [91–95]. Interestingly, women are often less dependent on nicotine then men [60, 96–98], they are less likely to be heavy smokers , and have lower concentrations of cotinine (a byproduct of nicotine). Notably, however, studies have consistently found that women have lower quit rates than men [61, 62, 97], have lower confidence in their ability to quit [98, 99], and often experience worse withdrawal symptoms during smoking cessation attempts [59, 97]. In fact, recent research suggests that the smoking rates for adolescent and adult women may actually be increasing .
Our study indicates that women are more likely than men to report emotional and sexual abuse and emotional neglect while men are more likely to report physical abuse and neglect. Literature has consistently indicated that women are more likely than men to report sexual abuse  and men are more likely than women to report physical abuse . The authors could find very little research that examined emotional abuse and neglect and physical neglect by sex; specifically research that did not contain the same data used in this study. The one study we did find indicated that women were significantly more likely than men to report emotional abuse and slightly more likely than men to report emotional neglect, although not statistically significantly so . This same study indicated that men were significantly more likely than women to report physical neglect, results consistent with our findings .
Much research has already examined potential biases and limitations of the ACE Study data. Research conducted by Felitti et al.  determined that respondent and nonrespondent groups were similar with regard to sociodemographic characteristics (e.g., percentages of women, mean years of education, and marital status), self-rated health, engagement in adverse health behaviors (e.g., smoking and other substance abuse), and presence of chronic diseases such as heart attack, stroke, chronic obstructive lung disease, hypertension, and diabetes. Edwards et al.  conducted research examining potential response bias and found that persons who did not participate in the ACE Study experienced childhood sexual abuse at the same rate as those who agreed to participate; research made possible by a dichotomous screening question about childhood sexual abuse in the health history survey. Moreover, those who participated in the study who reported sexual abuse had similar levels of current mental and physical health problems as those who did not participate and also reported sexual abuse . Test-retest reliability research conducted by Dube et al. , found that childhood sexual, physical, and emotional abuse, as well as forms of household dysfunction (i.e., mental illness in household, substance abuse in household, parental discord or divorce, incarcerated household member, and domestic violence), showed good Cohen’s Kappa agreement as defined by Fleiss  and Landis and Koch  (range = 0.46–0.86). Finally, while persons in the ACE Study are older, more educated, and less likely to smoke than the general population, ACE Study sexual and physical abuse estimates are similar to those derived from adult population-based surveys [108–110].
There are several limitation that warrants further examination. First, the ACE Study data are cross-sectional and do not collect specific information on temporality. Although most current literature suggests that the majority of psychiatric disorders associated with smoking occur prior to smoking initiation [29–44], other pathways have been posited (e.g., bidirectional association, common environmental and genetic factors for both, and smoking initiation prior to psychological distress) [111, 112]. Notably, in a study designed to specifically examine the stress-smoking relationship among adolescents, negative life events and negative affect were related to an increase in smoking over time, with no evidence of reverse causation . Additional longitudinal studies are needed to further clarify the relationships between ACEs, psychological distress, and smoking among adults. Second, there is undoubtedly more than one pathway that would lead an adolescent to smoking initiation (e.g., peer pressure). Moreover, there could be a cohort effect because participants in this study likely began to smoke at a time when smoking was more socially acceptable than it is now, and therefore the relative contribution of ACEs and psychological distress may increase or decrease as the rates of smoking decrease over time. Third, at the inception of the study, domestic violence was recognized to primarily occur against women. It is commonly known now that domestic violence occurs to both men and women in the household. Given this, our study has underestimated the prevalence of domestic violence in the household. Fourth, longitudinal follow-up studies of adults with documented childhood abuse suggested that retrospective reports of childhood abuse often underrepresented actual events [113–115]. However, in a recent study by Tourangeau and Yan , the authors indicate that respondents are less likely to underreport undesirable events and behaviors when the questions are self-administered and when the data are collected in private. Bias also may be introducted if there are differences in reporting retrospective information about childhood abuse by sex. In an article by Widom and Morris , among persons with a history of documented sexual abuse in childhood, fewer men than women later considered the event sexual abuse. Fifth, according to recent research, the joint effect of multiple ACEs on mental disorders are non-additive and often attenuate with age. This, combined with recall failure, often overestimates the effects of summary ACE scales . Given this, as was found in this study, one might not expect to see a dose–response relationship between number of ACEs and psychological distress. Further research is needed to determine an appropriate summary measure for retrospective studies. Sixth, it is not plausible that women would have more exposure to several of the ACEs (eg, household dysfunction) than men. This suggests that women are more sensitive to several of the ACE measures or are more willing to report them. Finally, psychological distress is a non-specific concept that can encompass everything from temporary negative emotion to chronic mental disorders. However, research suggests that the MCS is a good predictor of depressive disorders .
This study has important policy implications for public health approaches to smoking cessation. Despite increasingly stronger disincentives to smoking, including higher tobacco taxes and fewer places to smoke, the rate of smoking in the U.S. fell only slightly, from 20.9% in 2005 to 19.3% in 2010. At this slow rate of decline, by 2020 the adult smoking rate will only have fallen to about 17% . Given the strong association between ACEs and smoking, interventions targeted to trauma survivors may enhance the effectiveness of broader-based anti-smoking efforts.