Since the economic crisis, suicide mortality has risen sharply among Korean adolescents aged 15 to 19 years (i.e., an increase of 40.7% from 5.4 per 100,000 in 2001 to 7.6 per 100,000 in 2005) and is the second leading cause of death among this population.1 Suicide and suicidal behaviors (suicidal ideation and suicide attempts) among adolescents have become major public health concerns in Korea, and it is thus important to determine the risk factors associated with such behaviors. In recent years, early age of onset of alcohol drinking, cigarette smoking, and sexual intercourse (i.e., during childhood) have been identified as important predictors, beyond the multiple risk factors in adolescence previously noted in the literature,2-5 of later suicidal behaviors (during adolescence).6-13 With the exception of alcohol drinking, which is usually measured by self-reported age at first alcoholic drink excluding those imbibed for religious reasons,10,12,13 age at initiation of cigarette smoking and sexual intercourse has been conservatively measured in most previous studies (i.e., "regular/daily smoking or smoking a whole cigarette" and "unwanted/forced sexual intercourse").7,8,14 By using such methodologies, previous studies may have overestimated the contribution of early initiation of cigarette smoking and/or sexual intercourse to suicidal behaviors. Furthermore, according to Jessor's Problem Behavior Theory,15 alcohol drinking, cigarette smoking, and sexual intercourse have been highly correlated, even occurring simultaneously in some cases. Kim16 and Schmid, et al.17 reported that concur-rent users of alcohol and cigarettes were at higher risk for developing other problem behaviors (e.g., leaving home, delinquency, violating school rules) and substance use (e.g., stimulants/sleeping pills, marijuana, or butane gas/glue), compared to drinkers who did not smoke, smokers who did not drink, and abstainers (neither drinking nor smoking). However, little is known about the nature and extent of the health effects of early concurrent participation in problem behaviors, particularly in regard to later suicidal ideation and suicide attempts.

According to the Korean Ministry of Health and Welfare, the average reported age of drinking and smoking initiation decreased from 15.1 and 15.0 years in 1998 to 13.1 and 12.5 years in 2006, respectively; the prevalence of adolescent female smokers (12.8%) was about twice as high as that of adult female smokers (5.6%); and age at first experience of sexual intercourse also decreased.16,18 These findings have prompted an increased interest in the subsequent psychological status of such individuals. Nonetheless, despite the numerous factors placing this population at higher risk for problematic and suicidal behaviors, relatively few studies have been conducted in this domain.

Therefore, this study examined the associations between early initiation of alcohol drinking, cigarette smoking and sexual intercourse with suicidal ideation and suicide attempts. This study also determined the effects on each outcome of concurrent engagement in these problem behaviors, controlling for various confounding factors.

Table 1 shows the grade at first participation in alcohol drinking, cigarette smoking, and sexual intercourse by gender. Overall, the lifetime prevalence for alcohol drinking (61.9% of males vs. 61.1% of females, p = 0.034), cigarette smoking (32.0% vs. 22.1%, p < 0.001), and sexual intercourse (6.4% vs. 3.2%, p < 0.001) was higher in males than in females. In addition, more males than females reported consuming alcohol (19.9% of males vs. 15.2% of females, p < 0.001), smoking tobacco (13.2% vs. 7.9%, p < 0.001), and having sexual intercourse (1.1% vs. 0.4%, p < 0.001) before attending middle school (preteen initiators).

Table 2 summarizes the descriptive information for all variables by gender. In general, females showed a greater risk for suicidal ideation (26.8% of females vs. 18.1% of males, p < 0.001), suicide attempts (5.5% vs. 4.1%, p < 0.001), unhealthy weight control behavior (57.5% vs. 31.0%, p < 0.001), sleep dissatisfaction (43.6 % vs. 37.7%, p < 0.001), and depression (45.8% vs. 35.7%, p < 0.001) than did males. However, current alcohol drinking (26.9% of females vs. 30.0% of males, p < 0.001), cigarette smoking (7.9 of females vs. 14.0% of males, p < 0.001), and butane gas or glue sniffing (0.9% of females vs. 1.8% of males, p < 0.001) were more prevalent in boys than in girls.

Table 3 provides the gender specific percentage distributions and unadjusted odds ratios (95% confidential intervals) of suicidal ideation and suicide attempts by all key variables, including the three problem behaviors. Male and female adolescents who had already initiated these behaviors at any period (either teen or preteen initiators) were at greater risk for suicidal ideation and suicide attempts compared to non-initiators of alcohol drinking, cigarette smoking, or sexual intercourse. In particular, younger initiation of alcohol drinking, cigarette smoking, or sexual intercourse was correlated with greater likelihood of suicidal ideation and actual attempts. That is, male and female preteen initiators were at much greater risk for suicidal ideation and suicide attempts than were male and female teen initiators.

Table 4 presents the results of multivariate logistic regression analysis in the form of odds ratios and 95% confidence intervals for the associations among alcohol drinking, cigarette smoking, and sexual intercourse initiation with suicidal ideation and suicide attempts, after controlling for all covariates. As observed in the descriptive analysis, although the magnitudes of the associations of problem behaviors with suicidal behaviors were considerably attenuated when covariates were taken into account, their associations remained significant, except with regard to some male and female teen or preteen initiators. This finding indicated that preteen initiation of alcohol drinking, cigarette smoking, and sexual intercourse was independently and significantly associated with elevated likelihood of suicidal ideation and suicide attempts in both genders. Moreover, such associations were stronger among the preteen initiator group than among the teen initiator and non-initiator groups.

Table 5 displays the results of the multivariate analysis focusing on concurrent involvement in problem behaviors, and presenting logistic regression parameter estimates as odds ratios. Overall, the odds ratios increased as the frequency of concurrent engagement in problem behaviors increased during the same period of initiation. Interestingly, the odds ratios showed different patterns of associations between problem behaviors and suicidal behaviors according to when the former was initially experienced. Preteen initiators were at greater risk of suicidal ideation and suicide attempts than were teen initiators given concurrent participation at identical frequencies in problem behavior(s). Indeed, preteen and teen initiators with two or more problem behaviors appeared much more likely to demonstrate suicidal behaviors than teen initiators reporting the same frequency of problem behaviors, but less likely than preteen initiators alone.

Using a large nationally representative sample of Korean adolescents, we found that early initiation of each problem behavior (i.e., alcohol drinking, cigarette smoking, and sexual intercourse) was independently and significantly associated with an increased risk of suicidal ideation and suicide attempts, even when all covariates were controlled (e.g., sociodemographic characteristics, substance use, unhealthy behaviors, and psychological disorders). A similar pattern of correlations was revealed in both genders. This finding underscores that adolescent males and females whose initiation into alcohol use, tobacco use, or sexual intercourse occurred in childhood may be vulnerable to suicidal ideation and suicide attempts in adolescence. Understanding why those who have engaged in drinking, smoking, or sexual intercourse during their early lives go on to seriously consider and/or attempt suicide in later life is a complex task. It is possible that individuals with early exposure to these problem behaviors tend to be at higher risk not only for later alcohol abuse/dependence, nicotine dependence, unwanted pregnancy, abortion, and sexually transmitted diseases but also for later illicit drug abuse and antisocial characteristics, potentially leading them to consider and/or to attempt suicide.10,11,14,17,20 This finding is consistent with previous studies among U.S. high school students12,13 reporting that preteen alcohol use (i.e., before age 13), compared to abstinence during this period, was significantly associated with both suicidal ideation and suicide attempts, and that initiation of alcohol use during the teen years showed a similar but less significant association with suicide and suicidal behavior. Cho et al.9 also found a significant association between cigarette smoking and suicidal ideation among adolescent males and females in the U.S., but did not find a relationship between early alcohol drinking and suicidal ideation and suicide attempts in either gender. Differences between these data and our results may be attributable to differences between the studies in terms of the measures employed; in particular, early onset of alcohol drinking and cigarette smoking in the former study (e.g., the grade in which a young person first got drunk and started smoking regularly) tended to be more conservative than those in our study (e.g., the grade in which they first drank alcohol and smoked). Both problem behaviors were treated as categorical variables in this study, whereas as continuous variables in the previous study. We do not know whether the measurement of the sexual intercourse variable used in this study was construed as including forced sexual intercourse; however, if so, our results support those of previous studies showing that both male and female adolescents who have been forced to have sexual intercourse appeared more likely to engage in serious suicidal ideation.7,8 Since we measured onset variables less conservatively than did previous studies, our findings may have derived stronger implications than previous studies regarding the early initiation of alcohol use, cigarette smoking, and/or sexual intercourse on suicidal ideation and suicide attempts.

Another distinctive finding of this study emerged in regard to the association between the likelihood of suicidal ideation and suicide attempts and the time at which the problem behaviors were initially enacted and whether they were enacted concurrently. That is, the likelihood of both suicidal ideation and suicide attempts increased when these problem behaviors were enacted concurrently, particularly among preteens. This finding also provided further evidence for our first hypothesis, and indicated that concurrent participation in multiple problem behaviors had a synergistic effect on suicidal behaviors, and that this synergistic effect may be more detrimental to preteen initiators. Further investigation is necessary to identify the patterns underlying these issues.

This study had several limitations that should be considered in interpreting and generalizing these findings. First, our findings were based on self-reported data in the absence of corroboration from other sources. In addition, each variable (e.g., depression, suicidal ideation, and suicide attempts) was measured by a single item rather than by a scale consisting of multiple items; this may have resulted in reliability and validity problems. Nonetheless, most recent previous studies focusing on various racial and ethnic groups have used this self-reported single question method, and this trend has increased in recent years.4-7,12,13 Second, the main predictor variables (problem behaviors) used herein were addressed retrospectively. Moreover, these variables generally carry socio-cultural taboos for adolescent populations. Thus, it is possible that study participants did not answer accurately. In addition, our data were obtained from only middle and high school students, thereby excluding those who had dropped out of school and/or had a poor attendance record at the time of the survey, and we also omitted those who did not answer three primary questions (i.e., the onset of alcohol drinking, cigarette smoking and sexual intercourse) from the analysis, even though this excluded group might be characterized by disproportionately greater participation in problem and suicide-related behaviors.12 Thus, because of these biases, our conclusions may be understated. Third, because certain variables were not addressed in the KYRBWS data, this study could not consider several forms of psychopathology that have been directly or indirectly associated with suicidal behaviors, including adverse childhood experiences, aggressiveness, eating disorders, clinical symptoms, and social support and capital.2-4 Finally, the KYRBWS included information on onsets of the problem behaviors but did not include information on onsets of the outcomes. Consequently, we were unable to draw inferences about the causal pathways between problem behaviors and suicide-related behaviors among Korean adolescents. Therefore, future research should consider adopting a longitudinal design to examine the relationships among our focal issues with greater precision.

Despite these limitations, this study has several important strengths. The major strength of the present study was its use of a nationally representative sample and its high response rate, which may enable to generalize our findings to the actual population of Korean adolescents. Moreover, this study was the first to analyze measures of simultaneous alcohol use, cigarette smoking, and sexual intercourse, yielding indications that early initiations of problem behaviors are important predictors of later suicidal ideation and suicide attempts among Korean adolescent males and females. In particular, preteens were more susceptible than teens to the synergistic effects of combinations of these problem behaviors. Given recent trends and unlimited access to alcohol, tobacco, and internet and/or cyberspace sex in modern Korean society, it is not surprising that the age at which problem behaviors manifest continues to decrease. Consequently, the rates of adolescent suicide-related behaviors may continue to increase in the future. In particular, suicidal ideation and suicide attempts in adolescence have been identified as important predictors of suicidal ideation and suicide attempts in adulthood.21 Therefore, integrated preventive intervention programs targeting the concurrent initiation of alcohol use, cigarette smoking, and sexual intercourse, especially among preteens, should take precedence over those focused on each of these issues separately, thereby reducing the potential risks for subsequent suicidal ideation and suicide attempts.