The psychological characteristics of tobacco dependence in a rural area of Japan.

To investigate the psychological characteristics of smoking and tobacco dependence, a questionnaire survey was conducted of 2,051 consenting adults aged 21 to 65 years from the total population of a town in Akita Prefecture. Valid responses were received from 673 men and 818 women, for a total of 1,491 (72.7%). The questionnaires used were a tobacco dependence questionnaire, a personality questionnaire (the NEO Five-Factor Inventory: NEO-FFI), the WHO Subjective Well-being Inventory (SUBI), and the 12-Item General Health Questionnaire (GHQ-12). Subjects who responded that they had smoked daily for more than one month before the day of the survey were considered smokers. There were 400 male (59.8%) and 39 female (4.8%) smokers. Both male and female smokers were more extroverted than nonsmokers. Among smokers, there were 166 (41.5%) men and 19 (49%) women diagnosed as having a tobacco dependence of ICD-10. Among men, dependent smokers had a significantly higher score for Openness of the NEO-FFI and a lower score for Perceived III Health on the SUBI than nondependent smokers. Among smokers, 128 (32.0%) men and 13 (33%) women met the ICD-10 criteria for tobacco withdrawal. Among male smokers, those with tobacco withdrawal had a significantly higher score for Neuroticism and Openness on the NEO-FFI and a lower score for Negative Affect on the SUBI than those with no withdrawal. These results suggest that there is a relationship between tobacco dependence and personality or negative emotions.


The Psychological
Characteristcs of Tobacco Dependence in a Rural Area of Japan

Kimio Yoshimura
To investigate the psychological characteristics of smoking and tobacco dependence, a questionnaire survey was conducted of 2,051 consenting adults aged 21 to 65 years from the total population of a town in Akita Prefecture. Valid responses were received from 673 men and 818 women, for a total of 1,491 (72.7%). The questionnaires used were a tobacco dependence questionnaire, a personality questionnaire (the NEO Five-Factor Inventory: NEO-FFI), the WHO Subjective Well-being Inventory (SUBI), and the 12-Item General Health Questionnaire (GHQ-12). Subjects who responded that they had smoked daily for more than one month before the day of the survey were considered smokers. There were 400 male (59.8%) and 39 female (4.8%) smokers. Both male and female smokers were more extroverted than nonsmokers.
Among smokers, there were 166 (41.5%) men and 19 (49%) women diagnosed as having a tobacco dependence of ICD-10. Among men, dependent smokers had a significantly higher score for Openness of the NEO-FFI and a lower score for Perceived III Health on the SUBI than nondependent smokers. Among smokers, 128 (32.0%) men and 13 (33%) women met the ICD-10 criteria for tobacco withdrawal. Among male smokers, those with tobacco withdrawal had a significantly higher score for Neuroticism and Openness on the NEO-FFI and a lower score for Negative Affect on the SUBI than those with no withdrawal. These results suggest that there is a relationship between tobacco dependence and personality or negative emotions. Related Health Problems (ICD-10) 5). In the ICD-10, tobacco dependence (F17.2) is considered to be one type of drug dependence (Flx.2) to which the same diagnostic criteria apply as are used with other drug dependence such as alcohol, opioids, cannabis, and cocaine. The Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV)6), U.S. standards for the diagnosis of mental disorders, refer to "tobacco dependence" as "nicotine dependence" since the cause of tobacco dependence is the nicotine in the tobacco. With the ICD-10, drug dependence is diagnosed when any 3 of the following 6 items apply: (1) strong desire for drugs , (2) uncontrol, (3) withdrawal state, (4) tolerance, (5) preoccupation and, (6) use despite harmful consequences, as shown in Table 1. Thus, these diagnostic criteria, focusing on psychological dependence, or the psychopharmacological factors of drug dependence, are embodied on the behavioral level including subjective symptoms of physical dependence, that is, tolerance or withdrawal. Tobacco withdrawal is defined separately in F17.3.
Much research has been done on the mechanism of tobacco dependence, from the standpoints of pharmacology, epidemiology, behavioral genetics, psychology, and psychiatry. Clarifying the biological and social characteristics of smoking behavior would be useful in studying the mechanism of dependence. In particular, research on the relationship between smoking behavior or tobacco dependence and personality may provide clues as to why some people smoke and fall more easily into tobacco dependence.
Many studies have compared the personalities of smokers and non-smokers [7][8][9][10][11][12][13][14][15][16][17][18][19]. There have also been studies on the differences in personality between dependent smokers and nondependent smokers 20) or on the relationship between withdrawal and personality 2). However, there have been no studies to date on tobacco dependence using a questionnaire based on the five-factor model, which is part of recent personality theory. Therefore, the present study was undertaken with the objective of investigating smoking and tobacco dependence and its psychological characteristics. A survey was conducted of the general population of Nishime Town, Akita Prefecture, Japan, with ages ranging from 21 to 65 years. Psychometrical assess- modified as a self-administered questionnaire, it is thought that the validity of this questionnaire was kept almost as well as CIDI.
The NEO-FFI is a questionnaire widely used today to measure personality, developed by Costa and McCrae of the U.S. National Institute on Aging22) as. This is composed of 5 scales with 60 items, based on a five-factor model used to evaluate personality traits according to 5 scales. The 5 factors are Neuroticism (N), Extraversion (E), Openness (0), Agreeableness (A), and Conscientiousness (C). An explanation of each factor is given in Table 2. According to their research, a relative consistency exists among the 5 personality factors, and there is agreement between the personality a subject uses to rate himor herself and that rated by the people around, such as spouse or friends. Moreover, these characteristics of this questionnaire are seen to transcend different cultures, and not be greatly affected by the passage of time. A genetic component for each of the factors has also been observed. The validity of the Japanese version was examined by Yoshimura et al. 27).
The SUBI, a self-administered questionnaire of 40 items developed by the World Health Organization (WHO) 23), used to measure the degree of a person's physical, mental, and social well-being through individual experience. This questionnaire focuses on a person's overall health, not based simply on the person's emotions, but also on how well he or she recognizes their individual situations. It consists of 2 axes, Positive Affect and Negative Affect, and can measure not just the negative psychological state, but also the positive psychological side such as feelings of achievement or confidence. The SUBI is also comprised of the following 11 subscales: (1) General Well-Being -Positive Affect, (2) Expectation-Achievement Congruence, (3) Confidence in Coping, (4) Transcendence, (5) Family Group Support, (6) Social Support, (7)  The GHQ-12, developed by Goldberg of the U.K . 24), is the most general self-administered questionnaire for screening for nonpsychotic psychiatric illness. A higher score indicates a poorer condition. The validity of the Japanese version of this test was investigated by Fukunishi et al. 29).
In this study, those who responded that they had smoked every day for more than one month up to the day of the survey were considered smokers, and those who responded that they had not were considered non-smokers .
Comparisons were made between the groups of scores on each psychological scale using Analysis of Covariance (ANCOVA) adjusting for age. Analyses were performed by gender. Persons with missing values were excluded from each of the analyses. The significance level was set at p < 0 .05.

RESULTS
There were 400 male and 39 female smokers, and 273 male and 779 female nonsmokers. The proportions of smokers for males and females were therefore 59.4% and 4.8%, respectively, the smoking rate for women being far lower than that for men. Among both men and women, the smokers were significantly younger than the nonsmokers. Because there was such a large difference in the smoking rate between men and women, the psychological measure was investigated by gender. Table 4 shows the scores for each scale of the NEO-FFI of smokers and nonsmokers among men and women. Among men, the personality trait of Extraversion was significantly higher in smokers than in nonsmokers, and that of Openness significantly lower. Among women, Extraversion was significantly higher among smokers than nonsmokers, and Agreeableness significantly lower.
For subjective well-being, there was no significant difference in either Negative Affect or Positive Affect between male smokers and nonsmokers (Table 4). In the subscales, General Well-Being -Positive Affect was significantly lower. Among women, Negative Affect was significantly lower among smokers than nonsmokers. An investigation of each subscale showed that Perceived Ill Health and General Well-Being -Negative Affect were significantly lower.
No significant difference was seen in GHQ scores among men, but female smokers had a significantly higher score than nonsmokers.
There were 166 men (24.7%) and 19 women (2.3%) diagnosed as having a tobacco dependence. Two hundred and thirty-four men (34.8%) and 20 women (2.4%) were nondependent smokers. Accordingly, 41.5% of male smokers and 49% of female smokers had a tobacco dependence (Table 3). There was no sex difference in the percentages (p=0.40, by Fisher's exact test).
Regarding the 1st item of diagnostic criteria of ICD-10 tobacco dependence (strong desire for drugs) there was agreement by 245 men (63.5%) and 26 women (68%); on the 2nd item (uncontrol) by 266 men (68.9%) and 27 women (71%); on the 3rd item (withdrawal state) by 176 men (45.6%) and 16 women (42%); on the 5th item (preoccupation) by 2 men (0.5%) and 0 women; and on the 6th item (use despite harmful consequences) by 114 men (29.5%) and 11 women (29%). The 4th item (tolerance) was excluded from the CIDI because almost no tolerance results from tobacco dependence.   Table 5 shows a comparison of age, number of cigarettes per day, and the psychological scale scores between smokers with and without a tobacco dependence among men. Those with a tobacco dependence tended to smoke more cigarettes than smokers without a tobacco dependence. Openness was significantly higher as a personality trait. A logistic regression analysis was carried out with the score on each scale of the NEO-FF1 and age as the explanatory variables and either a dependent smoker or a nondependent smoker as the response variable. As a result, it was found that only the regression coefficient of Openness was significant (p=0.03), suggesting that the level of Openness is an independent risk factor of tobacco dependence among male smokers. On subjective well-being, the score for Perceived Ill Health of SUBI was significantly lower among men, showing that there were more physical complaints. No significant difference was seen in GHQ. Among women, no significant results were obtained in any scores.
The diagnostic criteria for tobacco withdrawal were fulfilled by 128 men (33.2%) among male smokers and 13 women (33%) among female smokers. Table 6 compares age, number of cigarettes per day, and the psychological scale scores between those who did and did not fulfill the criteria for tobacco withdrawal among male smokers. The scores for Neuroticism and Openness were significantly higher for those who fulfilled the diagnostic criteria for tobacco withdrawal than for those who did not. The same logistic regression analysis that was carried out for dependence was also carried out for withdrawal. The regression coefficients of Neuroticism and Openness were significant (p=0.04, 0.02, respectively), suggesting that the high score of each is an independent risk factor of tobacco withdrawal among male smokers. Negative Affect,

Inadequate
Mental Mastery, Perceived Ill Health, and Deficiency in Social Contacts of SUBI were all significantly low among male smokers with tobacco withdrawal. The GHQ score also tended to be high. No significant results were observed in any score among women.

DISCUSSION
This study revealed that smokers are more extroverted than nonsmokers among both men and women. Moreover, 41.5%   The result that smokers have a higher Extraversion score than nonsmokers agrees with past studies using NEO 19). In many investigations using other personality measures, extraversion was higher among smokers than nonsmokers 7-9, 11-15.17,15 For measures other than Extraversion in the present study, sex differences were recognized; among males smokers scored lower than nonsmokers in Openness, while females scored low in Agreeableness. According to the study by Kikuchi et al ., among males smokers tended to score lower in Openness and Neuroticism than nonsmokers, whereas no such differences were seen among females 19).
Madden et al. grouped twin subjects into 3 groups according to the severity of nicotine withdrawal and investigated its rela-tionship with the personality scales: Eysenck Personality Questionnaire (EPQ) 32) and Tridimensional Personality Questionnaire (TPQ) 33). They found that only Neuroticism of EPQ was significantly related. Harm Avoidance of TPQ tended to be high in those with moderate or severe withdrawal, and no relationships with other personality traits were reported 21). From a study using TPQ, Pomerleau et al. reported that Novelty Seeking and Reward Dependence were related to the start of smoking, and that the degree of dependence was related to Harm Avoidance 16). Breslau et al. also reported that the Neuroticism of the EPQ was related to tobacco dependence 20). Because Harm Avoidance of TPQ is closely related with Neuroticism of NEO-FFI or EPQ, it may be predicted that a difference would appear with regard to tobacco dependence or tobacco withdrawal for Neuroticism only in the NEO-FFI. However, in the present study, a difference was seen in Neuroticism for tobacco withdrawal, but no relation was found between tobacco dependence and Neuroticism.
In this study, a relationship was observed between Openness and both tobacco withdrawal and tobacco dependence. The concept of Openness appears on neither the EPQ nor TPQ. Since there are no past reports in which the relationship between tobacco dependence or withdrawal and personality was investigated using NEO, there are no past studies with which a direct comparison can be made. There have long been reports based on twin studies claiming a genetic influence on personality as well as on smoking behavior 34-38), but recently the relationships between genetic polymorphism and personality as well as smoking behavior have been clearly demonstrated as developing molecular biology. That is, a relationship between genetic polymorphism of the D2 dopamine receptor (DRD2) and nicotine addiction has been reported 37-41). A relationship between genetic polymorphism of the DRD2 and Novelty Seeking has also been indicated42). Openness of the NEO-FFI would seem to imply a wide range of interests and a tendency toward positive behavior; if it is considered to be closely related to Novelty Seeking on the TPQ, it may be that the present results occur due to polymorphism of the DRD2.
That perceived health was poor among those with a dependence suggests that there may be some people who cannot quit smoking even though they want to because of their poor physical condition. It may also be that the smoking of dependent smokers is the cause leading to their poor physical condition.
Given the limitations of the present study, ex-smokers, meaning those who had successfully quit smoking, were not considered. A comparison of ex-smokers and those with a tobacco dependence would perhaps reveal a clearer difference between those with and without a tobacco dependence. In addition, because ours was a cross-sectional survey in one specific region, it is not clear whether or not the present findings can be generalized. A future study in a different region or a follow-up survey of the same subjects would be beneficial.
Future studies should include demographic factors, social factors, comorbidity of mental symptoms, and genetic factors related to smoking '3). An elucidation of the causes of complex smoking behavior may also contribute to the improvement of anti-smoking or smoking prevention measures.

ACKNOWLEDGEMGEMENTS
The author would like to thank Prof. Masahiro Asai, Drs. Naohito Yamaguchi, Yutaka Ono, and Kazuo Suzuki for their kind comments and suggestions. This research was supported in part by a Grant-in-Aid for the Second Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health and Welfare, Japan.