Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, Williamson DF, Giovino GA. Adverse Childhood Experiences and Smoking During Adolescence and Adulthood. JAMA. 1999;282(17):1652-1658. doi:10.1001/jama.282.17.1652
Author Affiliations: National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, Ga (Drs Anda, Croft, Giles, Williamson, and Giovino); Department of Preventive Medicine, Southern California Permanente Medical Group (Kaiser Permanente), San Diego (Dr Felitti); Department of Pediatrics, Emory University School of Medicine, Atlanta (Dr Nordenberg).
Context In recent years, smoking among adolescents has increased and the decline
of adult smoking has slowed to nearly a halt; new insights into tobacco dependency
are needed to correct this situation. Long-term use of nicotine has been linked
with self-medicating efforts to cope with negative emotional, neurobiological,
and social effects of adverse childhood experiences.
Objective To assess the relationship between adverse childhood experiences and
5 smoking behaviors.
Design The ACE Study, a retrospective cohort survey including smoking and exposure
to 8 categories of adverse childhood experiences (emotional, physical, and
sexual abuse; a battered mother; parental separation or divorce; and growing
up with a substance-abusing, mentally ill, or incarcerated household member),
conducted from August to November 1995 and January to March 1996.
Setting A primary care clinic for adult members of a large health maintenance
organization in San Diego, Calif.
Participants A total of 9215 adults (4958 women and 4257 men with mean [SD] ages
of 55.3 [15.7] and 58.1 [14.5] years, respectively) who responded to a survey
questionnaire, which was mailed to all patients 1 week after a clinic visit.
Main Outcome Measures Smoking initiation by age 14 years or after age 18 years, and status
as ever, current, or heavy smoker.
Results At least 1 of 8 categories of adverse childhood experiences was reported
by 63% of respondents. After adjusting for age, sex, race, and education,
each category showed an increased risk for each smoking behavior, and these
risks were comparable for each category of adverse childhood experiences.
Compared with those reporting no adverse childhood experiences, persons reporting
5 or more categories had substantially higher risks of early smoking initiation
(odds ratio [OR], 5.4; 95% confidence interval [CI], 4.1-7.1), ever smoking
(OR, 3.1; 95% CI, 2.6-3.8), current smoking (OR, 2.1; 95% CI, 1.6-2.7), and
heavy smoking (OR, 2.8; 95% CI, 1.9-4.2). Each relationship between smoking
behavior and the number of adverse childhood experiences was strong and graded
(P<.001). For any given number of adverse childhood
experiences, recent problems with depressed affect were more common among
smokers than among nonsmokers.
Conclusions Smoking was strongly associated with adverse childhood experiences.
Primary prevention of adverse childhood experiences and improved treatment
of exposed children could reduce smoking among both adolescents and adults.
The dramatic rate of decline in the prevalence of cigarette smoking
among US adults that took place over the past 30 years has slowed and nearly
come to a halt.1,2 Furthermore,
past declines in smoking among US adolescents have actually reversed during
the 1990s,1,3 and about two thirds
of US adults who have ever smoked regularly have tried smoking by the age
of 18 years.4 These recent disturbing trends
in smoking have occurred amidst efforts to reduce access to cigarettes and
counter the effects of tobacco marketing, parent and sibling smoking, and
peer pressure to smoke.4- 9
Some of this reversal may also be due to exorbitant advertising and promotion
expenditures,10 especially for brands that
appeal to youth,11 and increasing exposure
to tobacco in the media.12 However, we have
an incomplete understanding of the reasons for this reversal. Further insight
into the basic underlying factors that lead to smoking during adolescence
and adulthood, be they depression,13- 16
anxiety,17 or social and developmental18 impairments, are needed.
We used data from the Adverse Childhood Experiences (ACE) Study19 to estimate the strength of the relationship between
adverse childhood experiences and 5 smoking behaviors: early smoking initiation,
smoking initiation as an adult, ever smoking, current smoking, and heavy smoking.
The adverse childhood experiences included the following: emotional, physical,
and sexual abuse; a battered mother; parental separation or divorce; and growing
up with substance-abusing, mentally ill, or incarcerated household members.
We included these childhood exposures because they can have a detrimental
effect on a child's emotional and social development and behavior,20,21 are common,19,22- 24
frequently co-occur,19 and represent health
or social problems of national importance.
This portion of the ACE Study is a retrospective cohort study analysis
designed to assess the effect of specific adverse childhood experiences on
adult health behaviors associated with the leading causes of morbidity and
mortality in the United States.19 The study
is being conducted among adult members of the Kaiser Permanente health maintenance
organization in San Diego, Calif. Each year, more than 45,000 adult members
undergo a standardized biopsychosocial medical examination at a primary care
clinic; 81% of Kaiser Permanente members in San Diego, aged 25 years or older,
who were continuously enrolled between 1992 and 1995, have been evaluated
at this clinic.
All 13,494 Kaiser Health Plan members who completed standardized medical
evaluations at the clinic in August through November 1995 or January through
March 1996 were eligible to participate in the ACE Study. Members examined
during December 1995 were not included because survey response rates are lower
during this holiday period.25 One week after
the clinic visit, these members were mailed a study questionnaire about health
behaviors and adverse childhood experiences. Questions from several disciplines
were used in the study questionnaire.22,23,26- 28
Questionnaires were returned by 9508 people (70.5%). Response rates did not
differ by sex, education, cigarette smoking behavior, or history of childhood
sexual abuse as recorded in the clinic's medical record; however, respondents
were slightly older (56.7 years vs 49.3 years) and more likely to be white
(84% vs 75%) than were nonrespondents.
We excluded 51 respondents with missing information about race, 34 with
missing information about educational attainment, and 208 with incomplete
information about smoking behaviors. Thus, the final study cohort included
97% of the respondents (9215/9508).
The ACE Study was approved by the institutional review boards of the
Southern California Permanente Medical Group, Emory University, and the Office
of Protection From Research Risks, National Institutes of Health.
All questions about adverse childhood experiences pertained to the respondents'
first 18 years of life. For questions from the Conflict Tactics Scale (CTS),26 the response categories were as follows: never, once
or twice, sometimes, often, or very often.
Verbal abuse was determined from answers to the following 2 questions
from the CTS26: (1) "How often did a parent,
stepparent, or adult living in your home swear at you, insult you, or put
you down?" and (2) "How often did a parent, stepparent, or adult living in
your home threaten to hit you or throw something at you, but didn't do it?"
Responses of "often" or "very often" to either item defined verbal abuse during
Two questions from the CTS were used to describe childhood physical
abuse26: "Sometimes parents or other adults
hurt children. While you were growing up, that is, in your first 18 years
of life, how often did a parent, stepparent, or adult living in your home:
(1) push, grab, slap, or throw something at you? or (2) hit you so hard that
you had marks or were injured?" Potential responses included never, once or
twice, sometimes, often, or very often. A respondent was defined as being
physically abused if the response was either "often" or "very often" to the
first question or "sometimes," "often," or "very often" to the second.
Four questions from Wyatt27 were adapted
to define contact sexual abuse during childhood: "Some people, while they
are growing up in their first 18 years of life, had a sexual experience with
an adult or someone at least 5 years older than themselves. These experiences
may have involved a relative, family friend, or stranger. During the first
18 years of life, did an adult, relative, family friend, or stranger ever
(1) touch or fondle your body in a sexual way, (2) have you touch their body
in a sexual way, (3) attempt to have any type of sexual intercourse with you
(oral, anal, or vaginal), or (4) actually have any type of sexual intercourse
with you (oral, anal, or vaginal)?" A "yes" response to any 1 of the 4 questions
defined a respondent as having experienced contact sexual abuse during childhood.
We used 4 questions from the CTS26 to
define childhood exposure to a battered mother. The questions were preceded
by the following question: "Sometimes physical blows occur between parents.
While you were growing up in your first 18 years of life, how often did your
father (or stepfather) or mother's boyfriend do any of these things to your
mother (or stepmother): (1) push, grab, slap, or throw something at her, (2)
kick, bite, hit her with a fist, or hit her with something hard, (3) repeatedly
hit her over at least a few minutes, or (4) threaten her with a knife or gun,
or use a knife or gun to hurt her?" A response of "sometimes," "often," or
"very often" to at least 1 of the first 2 questions or any response other
than "never" to at least 1 of the third and fourth questions defined a respondent
as having had a battered mother.
Two questions asked whether the respondent, during his or her childhood,
lived with a problem drinker or alcoholic24
or anyone who used street drugs. An affirmative response to either question
indicated childhood exposure to substance abuse in the household.
A respondent who said that during his or her childhood, anyone was depressed
or mentally ill or that anyone in the household had attempted suicide was
considered to have been exposed to mental illness in the household.
This was defined as a "yes" response to the question, "Were your parents
ever separated or divorced?"
If anyone in the household had gone to prison during the respondent's
childhood, this was defined as having childhood exposure to a household member
who was incarcerated.
We used the following standardized definitions for smoking behaviors28,29: early smoking
initiation as regularly smoking cigarettes by 14 years of age; ever smoked as having smoked at least 100 cigarettes; smoking
at the time of the survey as current smokers; and heavy smoking as currently smoking 20 or more cigarettes
per day. We also assessed smoking initiation at age 19 years or older to assess
adult smoking initiation to provide a subgroup of persons among whom we were
certain that all adverse childhood experiences antedated smoking initiation.
Respondents who reported that either parent smoked during the respondent's
childhood were considered to have a history of parental smoking.
Because we hypothesized that the relationship between current smoking
and adverse childhood experiences may be a result of the use of nicotine to
"self-medicate" affective disorders with the psychoactive actions of nicotine,30 we assessed the relationship between adverse childhood
experiences, current smoking, and depression during the past year. We used
a question from the Diagnostic Interview Schedule31
to assess recent problems with depressed affect: "Have you felt depressed
or sad much of the time in the past year?"
Persons with incomplete information about an adverse childhood experience
were considered not to have had that experience. This would likely result
in conservative estimates of the relationships between adverse childhood experiences
as persons who had potentially been exposed to an experience would always
be misclassified as unexposed; this type of misclassification would bias our
results toward the null.32 However, to assess
this potential effect, we repeated our analyses after excluding any respondent
with missing information on any one of the adverse childhood experiences.
Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were obtained
from logistic regression models that assessed the associations between each
category of adverse childhood experience and smoking behaviors. The number
of adverse childhood experiences was summed for each respondent (range, 0-8);
analyses were repeated with the summed score as an ordinal variable (0, 1,
2, 3, 4, or ≥5) or as 5 dichotomous variables (yes/no) with 0 experiences
as the referent. Covariates in all models included age, sex, race (other vs
white), and education (high school diploma, some college, or college graduate
vs less than high school).
To assess whether the relationship between the adverse childhood experiences
and smoking behaviors was explained primarily by either the genetic or role-modeling
influences of parental smoking or substance abuse (alcohol or illicit drugs)
in household members, we repeated separate logistic models stratified by the
presence or absence of a history of parental smoking and the presence or absence
of substance abuse in the household.
The study population included 4958 women and 4257 men. The mean (SD)
age was 55.3 (15.7) years for women and 58.1 (14.5) years for men. Seventy-seven
percent of women and 81% of men were white; 37% of women and 48% of men were
college graduates; and another 33% of women and 30% of men had some college
Nearly 70% of respondents reported a history of parental smoking (Table 1). Early smoking initiation was
more common among men; the prevalence of current smoking was similar for women
and men (Table 1). Because our
study population is better educated and older than the general population
and the prevalence of smoking tends to be lower in these groups, we adjusted
for differences in the demographics of our study population (educational attainment,
age, sex, and race) and the population of California using the 1995 census.
The adjusted prevalence of current smoking was 14.4% in the study sample,
which is similar to the estimate of 15.5% for the prevalence of current smoking
for all California residents in 199533; thus,
the apparently low prevalence of smoking in the study population can be accounted
for by its demographic characteristics.
With the exception of physical abuse, women were more likely than men
to report each category of adverse childhood experience (Table 1). Sixty-three percent of women and men reported at least
1 category of adverse childhood experience.
If a respondent was exposed to one of the adverse childhood experiences,
the probability of exposure to any other category of adverse childhood experience
was increased substantially (Table 2).
The median probability of exposure to any additional category given exposure
to a first was 85.5%; for any 2 additional categories, the median probability
The strengths of the relationships between each of the categories of
adverse childhood experiences with smoking initiation and current smoking
behavior were comparable (Table 3).
All categories of adverse childhood experiences were also significantly associated
with ever smoking and heavy smoking (P<.05; data
The relationship between the frequency of exposure to emotional abuse,
physical abuse, and maternal battery and early smoking initiation and current
smoking using the questions about the least severe forms of abuse or violence
from the CTS26 showed a positive graded relationship
between the frequency of exposure to these forms of violence and the prevalence
of these behaviors (P<.001; data not shown).
We performed a separate analysis in which we used a more conservative
definition of sexual abuse (abuse that occurred by 14 years of age or younger)
and excluded respondents whose age of initiation was earlier than the age
at which the sexual abuse first occurred. In this analysis, the prevalence
of early smoking initiation was 25.1% for persons who had been sexually abused
and 8.7% for those who had not (adjusted OR, 3.9; 95% CI, 3.2-4.7); the prevalence
of current smoking for those respondents who had been sexually abused was
14.4% and 8.2% for those who had not (adjusted OR, 2.7; 95% CI, 2.2-3.3).
Because the strength of the relationships between each of the adverse
childhood experiences and the smoking behaviors were comparable, we used a
simple additive approach to categorize the number of adverse childhood experiences.
We found strong graded relationships between the number of categories of adverse
childhood experiences and each smoking behavior (P<.001)
(Table 4). The relationship between
the number of categories of adverse childhood experiences and smoking behaviors
in logistic models that were stratified by either a history of parental smoking
or household substance abuse were similar in each stratum (data not shown).
When we repeated the analyses in Table 4 after excluding any person with missing information about an adverse
childhood experience, we found similar, strong, graded relationships between
adverse childhood experiences and each smoking behavior (data not shown).
The mean age of initiation among ever smokers for those with no adverse
childhood experiences was 20.9 years, whereas for those with all 8 experiences,
the mean age was 17.3 years. The relationship between age at initiation and
number of adverse childhood experiences was inverse and strongly graded (from
0 through 8 categories, ages were 20.9, 19.3, 19.0, 19.4, 18.6, 18.5, 17.4,
17.5, and 17.3 years, respectively; P<.001; multiple
We assessed the possibility that changing trends in social forces and
knowledge about the risks of smoking may have affected the relationship between
adverse childhood experiences and smoking behaviors. To do this, we performed
2 birth cohort analyses. The first analysis divided the subjects into those
born during 1902 and 1939 and those born between 1940 and 1979. We chose these
cohorts because those born during World War II or later were unlikely to have
been influenced by rationing of cigarettes by the armed forces and also lived
with increasingly intense pressures not to smoke or to quit. The second birth
cohort analysis divided the subjects into those born between 1902 and 1959
and 1960 and 1976; the younger birth cohort grew up after the surgeon general's
warning against smoking and also faced intense social pressures not to smoke.
In both analyses, in the younger birth cohorts the associations between adverse
childhood experiences and each smoking behavior tended to be stronger (data
The percentage of persons who reported feeling depressed or sad much
of the time in the past year increased as the number of adverse childhood
experiences increased. Furthermore, for any given number of adverse childhood
experiences, the percentage of respondents reporting problems with depressed
affect was always higher among smokers than among nonsmokers (Figure 1).
To our knowledge, this is the first study to assess the relationship
between a full range of common and interrelated adverse childhood experiences
and smoking behaviors. Few studies have assessed the association between childhood
abuse and smoking among adults34- 36
or adolescents.37- 39
Some studies have found associations of childhood abuse with substance and
alcohol abuse but only marginal associations with cigarette smoking.37,38 We found that the relationship between
the number of categories of adverse childhood experiences and each of the
smoking behaviors is strong and cumulative.
The associations that we report were similar when we stratified our
analyses by a history of parental smoking and household substance abuse. Thus,
the relationship between adverse childhood experiences and smoking behaviors
does not appear to be mediated primarily by genetic influences40- 43
from, or modeling of smoking by, parents who were smokers, alcoholics, or
illicit drug abusers. However, there may be unmeasured genetic components
to the relationships between adverse childhood experiences and smoking.43
Some of the exposures we studied, such as to household substance abuse,
seem to play a causal role in the occurrence of other exposures. Our data
indicate that if a person reports one of these adverse childhood experiences,
there is an 85% chance of experiencing a second, and a 70% chance of experiencing
a third. In light of these data, the historical tendency to focus on the effects
of single types of adverse childhood experience such as childhood sexual abuse
and the small number of studies that assessed the impact of more than 1 type
of abuse may be limiting.44- 49
This tendency leads to underestimation of the full exposure burden and hence
to underestimating the importance of adverse childhood experiences to important
health and social outcomes. Nonetheless, future analyses of the ACE Study
data are planned and data from other studies will be important to understand
the relationship of specific adverse childhood experiences to health outcomes.
Sexual abuse that occurred by 14 years of age and antedated age of smoking
initiation was associated with a 4-fold increase in smoking initiation. For
adult smoking behaviors, including smoking initiation at age 19 years or older,
current smoking, and heavy smoking, we could be certain that the childhood
experiences antedated the smoking behavior. Our data showing that the mean
age at smoking initiation was inversely related to the number of adverse childhood
experiences provide further support for a causal relationship between adverse
childhood experiences and smoking initiation.
We now come to a crucial question posed by our findings: How might persons
exposed to childhood adversity benefit from the use of nicotine? Nicotine
has demonstrable psychoactive benefits in the regulation of affect50; therefore, persons exposed to adverse childhood
experiences may benefit from using nicotine to regulate their mood.30,50,51 For such persons,
attempts to quit may remove nicotine as their pharmacological coping device
for the negative emotional, neurobiological, and social effects of adverse
childhood experiences. That is, nicotine appears to be a sufficiently effective
psychoactive agent that unconscious selection of its recurrent use could occur
in situations of chronic distress. This hypothesis is supported by our data
showing a strong, graded relationship between past-year depression, and that
for any given number of adverse childhood experiences, current smokers were
always more likely to have problems with depression.
Experiments with animals have shown that experiences that would be considered
adverse for human children are more likely to lead to self-administration
of nicotine52 and other drugs.53
Since young children are dependent on their parents or other household members
for their survival, the aversive stimuli they experience in abusive, violent,
or dysfunctional households are as inescapable as those administered to animals
in laboratory settings.
Because of the negative biological and emotional impact of adverse childhood
experiences, children who experience them may be victims on 4 levels. First,
they are victims of abuse, violence, and other dysfunction in their households.
Second, because of resulting problems with affect, socialization, and self-esteem
they may be more likely to fall prey to both peer pressure and the seductive
marketing practices of the tobacco industry, which spent $5.1 billion in tobacco
promotion and advertising in 1996.10 Third,
many states have passed legislation that fines or criminalizes children for
the purchase, possession, or use of tobacco.54
The number of states with such legislation increased from 32 in 1995 to 41
in 1998; the current maximum state penalty for minors is a fine of up to $1000,
and it is possible for fines to minors to exceed the fines to retailers who
sell tobacco to them (unpublished data, Centers for Disease Control and Prevention,
Office on Smoking and Health, State Tobacco Activities Tracking and Evaluation
System [STATE], 1998). Although some form of negative consequences for youth
who purchase or possess tobacco might be useful, the effectiveness of current
state sanctions is unproven and, in some cases, seem excessive. For youths
whose smoking may be a consequence of adverse childhood experiences, they
are victimized a third time by legislation that fines or criminalizes them
for responding to cigarette advertising and, perhaps, finding that nicotine
provides pharmacological relief from the effects of adverse childhood experiences.
Finally, as adults they are likely to become victims of diseases caused by
Because adverse childhood experiences are common and strongly associated
with smoking initiation, preventing their occurrence56
and early identification and treatment of children exposed to them may reduce
smoking initiation among adolescents. Current smokers who consciously or unconsciously
use nicotine as a pharmacological tool to alleviate the long-term emotional
and psycho-biological wounds of adverse childhood experiences may need special
assistance to help them quit. Such assistance includes recognition of the
use of nicotine to modulate problems with affect, treatment of the residua
of these adverse childhood experiences, and the use of nicotine replacement
therapy57 or antidepressant medications.58 These efforts could contribute substantially to the
reestablishment of the historical downward trends in smoking initiation and
smoking prevalence in the United States.