The ACE Study sample was composed of 9367 women (54%) and 7970 men (46%) with a mean age of 54.8 years (standard deviation [SD] = 15.7) among women and 57.5 years (SD = 14.6) among men. More than three-quarters of the participants (78.5% of the women and 81.6% of the men) described themselves as white; 34.5% of women and 45% of men were college graduates; another 37.5% of the women and 34% of the men had some college education.
The current smoking prevalence was 8.8%; men were slightly but not significantly more likely to be current smokers than were women (9.1% vs 8.6%, χ2 [1, 17,160] = 1.47, not significant). On average, smokers smoked 15 cigarettes per day (SD = 10.3). Of the 1518 current smokers in the sample, 51.2% (776) reported one or more of the tobacco-related illnesses or symptoms.
lists the prevalence of each of the selected disease conditions and symptoms by sex, of which one or more were reported by 43.6% (7554) of the sample. The most frequently reported condition was high blood pressure, reported by >27% of the sample. Lung cancer was the rarest condition reported, having an overall prevalence of 0.3%.
ACEs were common among participants (). Only 36.1% of individuals reported no ACEs. In contrast, 37.9% of all respondents reported ≥2 ACEs. For all ACEs except physical abuse, female respondents reported significantly higher prevalences than did men.
Prevalence of individual ACEs and ACE score by sex
To examine the relation between ACEs and smoking persistence, we performed a series of analyses among those participants who reported any of the diseases or symptoms listed in and whose smoking status we knew (n = 7483). Within this group, the smoking prevalence was somewhat higher (n = 776; 10.4%) than in the total ACE sample. In addition, respondents in this group were significantly more likely to report physical, sexual, or emotional abuse (29.2% vs 27.0%, χ2 [1, 14,643] = 3.97; 21.7% vs 19.6%, χ2 [1, 14,643] = 9.72; and 11.3% vs 9.8%, χ2 [1, 14,643] = 9.38, respectively) at the .05 level. However, they did not report statistically higher prevalences of the other ACEs.
We first constructed a logistic model with the ACE Score as an ordinal predictor, adjusting for sex and age () and current smoking as the dependent variable. The ORs and 95% CI for smoking are displayed in , under the “Separate Model” heading. The overall test for trend was significant (p < .001). Strong, graded relations were found between the ACE Score and the likelihood of continuing to smoke despite having health problems that contraindicated smoking. The adjusted likelihood of being a current, persistent smoker rose from 1.08 in individuals reporting one ACE to 1.69 in individuals reporting ≥4 ACEs. The prevalence of persistent smoking rose in a dose–response fashion as the number of reported ACEs increased, rising from 7.8% in participants with no ACEs to 16.6% in those reporting ≥4 ACEs.
ACEs and the prevalence and risk
(adjusted OR) of smoking among adults with smoking-related diseases or symptoms
Next, we tested the relation between past or current depression and persistent smoking among those with smoking-related diseases and conditions. The adjusted odds ratio (OR) is also shown in in the column labeled “Separate Model.” Those who reported past or current depressed affect were 1.59 times more likely to be persistent smokers than those who did not affirm past or current depression (p < .001).
Finally, we constructed a model with both the ACE Score and past or current depression as predictors. These results are shown in under the column labeled “Single Model.” The χ-square value for the difference in the log likelihood ratios between the two models was significant (χ2 = 86.71, degrees of freedom = 1; p < .001); however, the addition of this variable only slightly attenuated the relation between the ACE Score and the odds of persistent smoking.
We also examined the ACE Score as a predictor of the number of smoking-related diseases and symptoms reported among smokers. We performed an analysis of variance with the number of smoking-related symptoms and diseases as the dependent variable (range, 0 to 10), controlling for age and sex. The ACE Score was statistically significant (F [1, 4] = 4.79; p < .001). A similar dose–response pattern of results was noted, wherein the average number of smoking-related symptoms and diseases increased as the number of reported ACEs rose ().
Average number of diseases/symptoms by adverse childhood experiences score (ACE), adjusted for age and sex.
Strong, graded relations were found between the ACE Score and the likelihood of continuing to smoke …