Abstract

Background

Sexual behaviors, particularly risky sexual behavior, has become a serious public health concern among adolescents worldwide, presenting a substantial obstacle to the prevention of sexually transmitted infections, including human immunodeficiency virus (HIV). However, there is limited research using consistent and standardized methodology to examine associations between tobacco and alcohol use frequency and both total and risky sexual behaviors among adolescents. We aimed to examine the association between tobacco and/or alcohol use with both total and risky sexual behaviors among adolescents worldwide.

Methods

Data were collected from the Global School-based Student Health Survey, which comprised 211,847 adolescents aged 12鈥�17 years from 59 countries. The frequency of tobacco or alcohol use during the past 30 days was categorized as 0, 1鈥�2, 3鈥�5, 6鈥�9, or 鈮モ��10 days. Tobacco and alcohol use were also categorized as non-use, tobacco use alone, alcohol use alone, and combined use. Multi-variable logistic regression analysis was used to examine both the independent and combined associations of tobacco and alcohol use with total and risky sexual behaviors.

Results

Compared with no tobacco use, the odds ratio of engaging in sexual intercourse increased with the frequency of tobacco use from 1 to 2 days to 鈮モ��10 days (total: 2.03 [95% confidence interval 1.47鈥�2.81] to 3.98[2.63鈥�6.03]; risky: 2.43[1.75鈥�3.38] to 4.21[3.26鈥�5.42]), as well as with the frequency of alcohol use. Overall, combined users had greater likelihood of both total and risky sexual behaviors than tobacco users alone, alcohol users alone, and non-users. Similarly, the association between risky sexual behaviors and tobacco use alone was more pronounced among adolescent girls (vs. adolescent boys), as were those of risky sexual behaviors with alcohol use alone among younger adolescents aged 12鈥�14 years (vs. aged 15鈥�17 years) and with tobacco and/or alcohol use among adolescents in the Western Pacific region (vs. Regions of Africa and Americas).

Conclusions

Our findings suggest independent and combined associations between tobacco and/or alcohol use with sexual behaviors among adolescents, with variations across age, sex, and WHO region.

Background

Sexual behaviors (ever had sexual behaviors) among adolescents, particularly risky sexual behaviors (i.e., early initiation of sexual behaviors, no condom use, and having multiple sexual partners), remains a public health concern worldwide. Data from the Global School-based Student Health Survey (GSHS) conducted in 69 low- and middle-income countries (LMICs) between 2003 and 2017 showed that 6.9% of young adolescents aged 12鈥�15 years reported ever engaging in sexual intercourse, among which 52% reported having multiple sexual partners and 41.9% reported not using a condom [1]. Sexual behaviors among adolescence has been associated with various negative outcomes among adolescents, including emotional distress, behavioral problems (e.g., suicide attempts [2], rule-breaking behavior, and aggressive behavior [3]), and unintended pregnancy [4]. Unprotected sexual behaviors poses a heightened risk of contracting sexually transmitted infections, including human immunodeficiency virus (HIV), particularly in LMICs [5,6,7]. It has been estimated that among 1.2听million adolescents aged 15鈥�19 years with HIV, 43% of male and 65% of female adolescents were infected by horizontal transmission, particularly through sexual behaviors [7]. In addition, according to reports from United Nations International Children鈥檚 Emergency Fund, around 15听million adolescent girls aged between 15 and 19 have experienced forced sexual behavior or other forms of sexual violence in their lifetime worldwide [8]. Therefore, implementing well-targeted public health interventions aimed at promoting safe sexual practices among adolescents holds substantial implications for the improvement of sexual and reproductive health.

Emerging evidence suggests that unhealthy behaviors, including tobacco and alcohol use, are associated with sexual behaviors among adolescents and young adults [9,10,11]. Meta-analyses conducted by Cho et al. [9] have indicated a positive association between alcohol use and total sexual behaviors, early sexual initiation, no condom use, and having multiple sexual partners among adolescents and young adults. However, these previous meta-analyses have identified significant between-study heterogeneity and publication bias. Additionally, most participants in the included studies were from high-income countries, and data from different studies using a standardized methodology and unified definition of variables are limited, making a cross-country or regional comparisons difficult. Moreover, the frequency of alcohol use and potentially confounding variables such as cannabis use [12] and dietary habits [13, 14] were not taken into account, which limits the development of more practical and accurate guidelines and interventions for distinguishing high-risk adolescents from those with low to moderate risk. Therefore, the findings of previous meta-analyses should be interpreted with caution, and future research should verify previous results using a larger sample size, a detailed classification of tobacco and alcohol use, the same methodology, and a comprehensive examination of potential confounding factors.

Using data from the GSHS between 2009 and 2015, Kushal et al. observed that tobacco use and alcohol use were identified as risk factors associated with early sexual initiation among adolescents aged 12鈥�15 years [11]. However, it is important to note that their study only classified adolescents as either users or non-users of tobacco and alcohol, without considering a 鈥渄ose-dependent鈥� association. Furthermore, the existing literature on the association between tobacco use and total and risky sexual behaviors has yielded inconsistent findings [15,16,17,18,19], which need further confirmation. Moreover, previous studies cannot determine whether tobacco use alone or alcohol use alone independently contributes to total and risky sexual behaviors among adolescents, and their combined effect also remains unknown.

Therefore, we used the latest nationally representative data from the GSHS between 2009 and 2019 to examine the independent association of tobacco and alcohol use frequency and their combined association with total sexual behaviors and risky sexual behaviors including early sexual initiation, having multiple sexual partners, and lack of condom use among adolescents aged 12鈥�17 years in 59 countries and territories (hereafter 鈥渃ountries鈥�).

Methods

Data sources

We used the most recent publicly available data from the GSHS surveys between 2009 and 2019 in 59 countries to assess the association between tobacco and/or alcohol use and sexual behaviors among adolescents aged 12鈥�17 years. In cases where a country had conducted more than one GSHS surveys between 2009 and 2019, we used the data from the last survey.

The GSHS is an ongoing, cross-sectional school-based survey that uses the same study methodology, validated questionnaire, and standardized two-stage cluster sampling strategy across all included countries. The GSHS is supported by the U.S. Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) in collaboration with the United Nations International Children鈥檚 Emergency Fund; United Nations Educational; Scientific and Cultural Organization; and the Joint United Nations Programme on HIV/AIDS. In the first stage of sampling, schools were selected at random using a probability-proportionate-to-size method. In the second stage, classrooms were randomly selected from the chosen schools. All selected students were invited to voluntarily complete a standardized and anonymous questionnaire. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline for cross-sectional studies. Details on the GSHS can be accessed on the websites of the US CDC and WHO [20, 21].

Ethical statement

The GSHS in each participating country was approved by an institutional ethics review committee or a national government administrative body, and informed verbal or written consent was obtained from the parents or guardians of all respondent.

Definition of sexual behaviors

To assess total sexual behaviors, respondents were asked: 鈥淗ave you ever had sexual behaviors?鈥� with response options of yes or no. To assess risky sexual behaviors, the following questions were used [22]: (1) 鈥淗ow old were you when you had sexual behaviors for the first time?鈥� with response options of never, age 11 years or younger, 12 years, 13 years, 14 years, 15 years, and age 16 or 17 years. We defined early sexual initiation as first having sexual intercourse at 14 years of age or younger [9, 11]. (2) 鈥淒uring your lifetime, with how many people have you had sexual intercourse?鈥� with response options of never, 1 person, 2, 3, 4, 5, and 6 or more people. We defined multiple sexual partners as having two or more partners. (3) 鈥淭he last time you had sexual intercourse, did you or your partner use a condom?鈥� with response options of never had sexual intercourse, yes, and no. Adolescents who responded 鈥渘o鈥� were categorized as non-users of condoms. We defined risky sexual behaviors as at least one of the above three risky behaviors (i.e., early initiation of sexual intercourse, no condom use, and having multiple sexual partners).

Definition of tobacco and alcohol use

The frequency of tobacco use was assessed using questions on cigarette use and any tobacco products other than cigarettes [22] : 鈥淒uring the past 30 days, on how many days did you smoke cigarettes?鈥� with response options of 0, 1 or 2, 3鈥�5, 6鈥�9, 10鈥�19,20鈥�29 days, and every day. A similar question was asked regarding tobacco products other than cigarettes. Tobacco use was defined as using cigarettes or any other tobacco products on at least one day during the previous 30 days. The frequency of alcohol use was measured using the question [22] : 鈥淒uring the past 30 days, on how many days did you have at least one drink containing alcohol?鈥� with response options of 0, 1 or 2, 3鈥�5, 6鈥�9, 10鈥�19,20鈥�29 days, and every day. The definition of alcohol use was analogous to that of tobacco use.

Covariates

The covariates considered in this study include age, sex, exposure to secondhand smoke, drug use (including marijuana, amphetamines, cocaine, and inhalants), intake of fast food, soft drinks, and fruit/vegetables, survey year, and World Bank income in the survey year for each country [23]. In certain countries, the questionnaire provided specific examples of fast foods and soft drinks, the details of which can be accessed in the WHO GSHS data repository [21].

Statistical analysis

We used primary sampling units, strata, and original sampling weights to calculate the prevalence of tobacco use, alcohol use, sexual behaviors, and risky sexual behaviors in each country. Overall and subgroup proportions of sexual behaviors and risky sexual behaviors were calculated by recalculating the sampling weights based on the sample size in each country. We used a linear trend test to estimate the trends in the proportion of sexual intercourse and risky sexual behaviors among adolescents with increased frequency of tobacco and alcohol use. Multivariable logistic regression models were used to examine the odds ratio (OR) and 95% confidence interval (CI) of independent and combined associations of tobacco and alcohol use with sexual intercourse and risky sexual behaviors adjusted for age, sex, secondhand smoke exposure, drug use, intake of fast food, soft drinks, and fruit/vegetables, survey year, and World Bank income. In addition, we performed subgroup analyses based on age (12鈥�14 years and 15鈥�17 years), sex (males and females), World Bank income (low-income, lower-middle income, upper-middle income, and high-income levels), and WHO regions (Africa, Americas, South-East Asia, and Western Pacific regions). None-overlapping 95% CIs were considered to indicate statistical differences between groups. All statistical analyses were conducted using SPSS 16.0. A P鈥塿alue鈥�<鈥�0.05 indicated statistically significance.

Results

Participant characteristics

A total of 211,847 adolescents (males: 50.7%) aged 12鈥�17 years from 59 countries across five WHO regions (i.e., 11 in Africa, 26 in America, 1 in Eastern Mediterranean, 6 in South-East Asia, and 15 in Western Pacific) including 9 low-income countries, 15 lower-middle income countries, 17 higher-middle income countries, and 18 high-income countries were enrolled in this study. Overall, the prevalence of current tobacco use was 9.3%, ranging from 2.6% in Mozambique to 51.1% in Tokelau. The prevalence of current alcohol use was 15.9%, ranging from 1.7% in Bangladesh to 53.9% in Argentina. The prevalence of sexual intercourse was 18.6%, ranging from 5.3% in Indonesia to 50.7% in Mozambique. The prevalence of risky sexual behaviors was 11.6%, ranging from 1.6% in Cambodia to74.4% in Uruguay (Table S1).

Change in proportions of total and risky sexual behaviors with frequency of tobacco or alcohol use

The proportion of sexual intercourse showed an upward trend from 0 days of tobacco use (15.2%) to 鈮モ��10 days of tobacco use (62.6%), regardless of sex, age, World Bank income, and WHO regions (all P鈥塮or trend鈥�<鈥�0.001, Fig.听1A and Table S2). Similarly, the proportion of sexual intercourse showed an upward trend from 0 days of alcohol use (14.6%) to 鈮モ��10 days of alcohol use (62.4%), regardless of sex, age, World Bank income, and WHO regions (all P鈥塮or trend鈥�<鈥�0.001, Fig.听1A and Table S3). Similar trends were found for risky sexual behaviors (except for total, adolescent boys, and adolescents from Lower-middle income countries of tobacco use from 6 to 9 days onwards during the past 30 days, Fig.听1B and Tables S4-5) and specific risky sexual behaviors including early initiation of sexual intercourse, no condom use, and multiple sexual partners (Tables S6-7).

Fig. 1

Proportions (%) of sexual intercourse and risky sexual behavior among adolescents by tobacco and alcohol use days during the past 30 days. (A) sexual intercourse; (B) risky sexual behaviors

Comparing proportions of total and risky sexual behaviors among users of both tobacco and alcohol, alcohol users alone, tobacco users alone, and none users

The proportion of sexual intercourse was highest among adolescents who used both tobacco and alcohol (56.6%), followed by those who used tobacco alone (36.8%), alcohol alone (28.8%), and those who used neither tobacco nor alcohol (13.0%, Table S8). Similarly, the proportion of risky sexual behaviors and specific risky sexual behaviors were also highest among adolescents who used both tobacco and alcohol and lowest among those who used neither tobacco nor alcohol, while no differences were found between tobacco alone and alcohol alone in most subgroup (Tables S9-10).

Change in odds of total and risky sexual behaviors with the frequency of tobacco or alcohol use

The OR values for sexual intercourse increased with the frequency of tobacco use during the past 30 days (ref. 0 days; 1鈥�2 days: 2.03, 95% confidence interval [CI] 1.47鈥�2.81; 3鈥�5 days: 2.82, 2.16鈥�3.66; 6鈥�9 days: 2.94, 2.06鈥�4.20; 鈮�10 days: 3.98, 2.63鈥�6.03), after adjusting for potential confounding factors. Subgroup analyses by sex (it should be noted that the highest OR was 4.19, 2.91鈥�6.05 in the group of tobacco use with 3鈥�5 days), age group, World Bank income group, and WHO region (except for the region of Americans, P鈥塮or trend was 0.18) showed similar trends (Table听1). The OR for sexual intercourse increased with the frequency of alcohol use during the past 30 days (ref. 0 days; 1鈥�2 days: 1.84, 95% CI 1.59鈥�2.13; 3鈥�5 days: 1.79, 1.38鈥�2.32; 6鈥�9 days: 2.48, 1.81鈥�3.39; 鈮�10 days: 4.01, 3.01鈥�5.34) after adjusting for potential confounding factors. Subgroup analyses by sex, age group, World Bank income group, and WHO region showed similar patterns (Table听2). Similar results were found for risky sexual behaviors (Tables S11-12), while the results for specific risky sexual behaviors, including early initiation of sexual intercourse, no condom use, and multiple sexual partners showed a different trends (e.g., an inverted U relationship for the specific risky sexual behaviors and tobacco use in most subgroup ) (Tables S13-14).

Table 1 Association between tobacco use and sexual intercourse among adolescents by sex, age group, World Bank income group, and WHO region

Table 2 Association between alcohol use and sexual intercourse among adolescents by sex, age group, World Bank income group, and WHO region

Comparing odds of total and risky sexual behaviors among users of both tobacco and alcohol, alcohol users alone, tobacco users alone, and none users

Compared with adolescents who used neither tobacco nor alcohol, those who used both tobacco and alcohol (OR 5.66, 95%CI 4.61鈥�6.96) had a greater likelihood of having sexual behaviors than those who used tobacco alone (3.24, 2.21鈥�4.75) or alcohol alone (2.37, 2.06鈥�2.73). Subgroup analyses by sex, age group, World Bank income group (except for lower-middle income countries with highest OR values for tobacco use alone), and WHO region showed similar results and no differences were found between groups (Fig.听2). Compared with adolescents who used neither tobacco nor alcohol, those who used both tobacco and alcohol (OR 5.98, 95%CI 4.79鈥�7.46) had a greater likelihood of having risky sexual behaviors than those who used alcohol alone (2.66, 2.32鈥�3.05) or tobacco alone (2.40, 1.66鈥�3.49) (Fig.听3). Subgroup analyses suggested that adolescent girls who used tobacco alone, younger adolescents aged 12鈥�14 years who used alcohol alone, and adolescents from the Western Pacific region who used tobacco alone, alcohol alone, or both substances tended to have greater odds of risky sexual behaviors than adolescent boys, older adolescents aged 15鈥�17 years, and those from other WHO region, respectively (Fig.听3). Similar results were found for specific risky sexual behaviors (Fig S1).

Fig. 2

Association of tobacco use alone, alcohol use alone, or combined use of both substances with sexual behaviors among adolescents by sex, age group, World Bank income group, and WHO region

Fig. 3

Association of tobacco use alone, alcohol use alone, or combined use of both substances with risky sexual behavior among adolescents by sex, age group, World Bank income group, and WHO region

Discussion

To the best of our knowledge, this is the first study to evaluate the global association between tobacco and alcohol use and both total and risky sexual behaviors among adolescents from 59 countries. We found that the proportion of sexual intercourse increased with higher frequencies of tobacco or alcohol use during the past 30 days. Furthermore, we found a 鈥渄ose-dependent鈥� relationship, where in the odds of total and risky sexual behaviors accelerated with an increase in tobacco or alcohol use. In addition, adolescents who reported concurrent use of both tobacco and alcohol had higher odds of engaging in both total and risky sexual behaviors than tobacco users alone and alcohol users alone. Furthermore, specific demographic and regional trends emerged, with adolescent girls who used tobacco alone, younger adolescents aged 12鈥�14 years who used alcohol alone, and adolescents from the Western Pacific region who used tobacco alone, alcohol alone, or both substances were more likely to engage in risky sexual behaviors.

Our study demonstrated an overall prevalence of tobacco use among adolescents at 9.3%, with a significant difference found across countries and WHO regions. Notably, the prevalence was lowest in Mozambique (2.6%) and highest in Tokelau (51.0%). These findings are consistent with the previous research, such as data from the 2006鈥�2013 GSHS across 68 countries, which suggested an overall prevalence of tobacco use was 13.6%, with the lowest rate in Tajikistan (2.8%) and the highest in Samoa (44.7%) [24]. Data from Health Behaviour in School-aged Children also showed similar results [25]. The main reason for the differences in smoking prevalence among countries might be due to the fact that some countries have not yet ratified the WHO Framework Convention on Tobacco Control (i.e., Tokelau) [26]. We also found that an overall prevalence of alcohol use among adolescents was 15.9%, with the lowest rate in Bangladesh (1.7%) and the highest in Argentina (53.9%). Data from the 1999鈥�2019 European School Survey Project on Alcohol and other Drugs (ESPAD) in 26 countries showed similar findings [27]. Differences in the prevalence of alcohol use across countries might be owing to differences in national alcohol prohibition policies. It is clear that countries with effective implementation of policies on underage drinking (e.g., prohibition of sales to adolescents) and countries with low social acceptance of underage drinking (i.e., Mozambique, Bangladesh) have lower rates of drinking [28]. These findings emphasize the urgent need to strengthen and implement public policies and programs to reduce tobacco and alcohol use among adolescents worldwide.

As far as we can ascertain, this is the first study to assess the global proportion of increased total and risky sexual behaviors according to more frequent use of tobacco or alcohol among adolescents. This addresses a limitation in previous studies conducted in specific countries, which predominantly used a binary classification of tobacco or alcohol use [9,10,11, 15, 16, 18, 29,30,31,32,33]. It has been previously reported that alcohol use is associated with sexual intercourse among adolescents [9, 34, 35]. Similar to two meta-analyses on alcohol use and sexual intercourse [34, 35], a recent meta-analysis involving 465,595 adolescents and young adults showed that alcohol use was associated with increased odds of risky sexual behaviors including early sexual initiation (OR 1.96, 95% CI 1.64鈥�2.35), no condom use (1.23, 1.11鈥�1.35), and multiple sexual partners (1.72, 1.53鈥�1.95) [9]. However, it is important to acknowledge that these meta-analyses face potential challenges in the form of high between-study heterogeneity and significant publication bias. Furthermore, it is worth noting that previous meta-analyses primarily categorized alcohol use as users versus non-users, and limited studies have focused on the 鈥渄ose-dependent鈥� association of alcohol use with sexual intercourse. Although one study based on four Pacific Island countries showed a gradient association between alcohol use and sexual intercourse, it is crucial to consider that the sample size was small, particularly for specific risky sexual behaviors [36]. In this study, we found that alcohol use was associated with increased odds of sexual intercourse, as well as risky sexual behaviors on a global scale. In addition, we found that there was a 鈥渄ose-dependent鈥� association from 0 days of use to 鈮モ��10 days of use. Further longitudinal cohort studies are required to validate the causality of these associations, as existing evidence shows that early sexual intercourse in adolescence is associated with substances misuse in adulthood [37].

Prior research has yielded inconclusive results regarding the association between tobacco use and sexual intercourse or risky sexual behavior among adolescents [10, 11, 15, 16, 19, 29,30,31,32,33, 36, 38, 39]. For example, a representative cross-sectional study in Ghana, involving 1,195 adolescents aged 12鈥�18 years demonstrated that tobacco use was significantly associated with increased odds of total sexual behaviors and multiple sexual partners [15]. Another cross-sectional study conducted with a sample of 2,668 senior high school students in China showed a positive association between cigarette use and sexual intercourse without condom use [16]. However, that study did not find a significant association between cigarette use and having multiple sexual partners [16]. In addition, a study based on genetically informative discordant twin design, including 3,400 adult same-sex twins in Australia, showed that smoking did not serve as a predictor of sexual intercourse onsets in adolescents [17]. Several previous studies involving adolescents from Ghana, Canada, and the United States also failed to identify a significant association between tobacco use and sexual intercourse without condom use [18, 19, 40]. The discrepancy might be owing to differences in population characteristics, age, sample sizes, and study designs. In this study, we first used global data collected using the same study methodology in 59 countries and found that tobacco use was significantly associated with total sexual intercourse and risky sexual behaviors, including early sexual initiation, sexual intercourse without condom use, and multiple sexual partners with a 鈥渄ose-dependent鈥� pattern, irrespective of age, sex, World Bank income, and WHO region. Additionally, we not only found an independent association of sexual intercourse with tobacco use alone or alcohol use alone but also observed that the odds for total and risky sexual behaviors among both tobacco and alcohol users were higher than those among tobacco users alone and alcohol users alone. The potential mechanisms might be explained by the detrimental effects of alcohol and tobacco on adolescents, which may increase psychological and behavioral problems such as depression and anxiety, affecting their sexual decision-making and rendering adolescents more susceptible to impulsive behaviors such as risky sexual behaviors [41, 42]. Our findings suggest that even a low frequency of tobacco use or alcohol use may contribute to total and risky sexual behaviors among adolescents, particularly for combined users, which provides valuable evidence for the development of effective strategies aimed at preventing risky sexual behaviors among adolescents worldwide.

Data from the 2003鈥�2017 GSHS in 69 countries suggested that the global prevalence of ever engaging sexual behaviors among adolescents aged 12鈥�15 years was higher among boys than girls (10.0% vs. 4.2%) [1]. Consistent with previous studies among adolescents from Brazil and the United States [10, 31], we also found that although adolescent boys experienced more sexual behaviors than adolescent girls, girls with tobacco use alone were more strongly associated with risky sexual behaviors including early sexual behaviors, no condom use, and multiple sexual partners (the results also apply to the both use). It is possible that female adolescents who engage in tobacco use are more prone to having more opportunities for engaging in risky sexual behaviors and associating with friends and peers of the opposite sex who encourage such behavior [43]. We additionally found that younger adolescents (12鈥�14 years old) with alcohol use alone were more strongly associated with risky sexual behaviors, including early sexual behaviors (the results also apply to the both use) and multiple sexual partners, in comparison with older adolescents (15鈥�17 years old). This could potentially be explained by the fact that compared with older adolescents, younger adolescents with alcohol use might lack mature emotional and cognitive abilities, which clearly places the sexual and reproductive health among younger adolescents (aged 10鈥�14 years) at stake [44]. Consistent with the findings in previous meta-analyses [9, 34], the strength of the association between alcohol use alone and risky sexual behavior did not differ between males and females. Therefore, implementing a population strategy that restricts access to tobacco and alcohol for all adolescents, along with a targeted approach that offers age- and sex-specific education, may prove more effective in mitigating risky sexual activity among this age group.

In this study, we found that the association of risky sexual behaviors with tobacco use alone, alcohol use alone, or combined use of both substances was notably significant in the Western Pacific region when compared to Africa, and the Americas. This might be explained by the fact that adolescents in certain Western-Pacific countries exhibit more permissive attitudes towards male sexual infidelity [45], have limited access to sexual and reproductive health [46], and experience increased obstacles to modern contraceptive use [47,48,49]. These findings suggest that implementing enhanced strategies for promoting sexual and reproductive health; reducing health disparities; providing access to condom dispensers; improving sexual perception, culture, and values; and educating adolescents on healthy sexual attitudes, especially those who engage in tobacco and/or alcohol use in the Western Pacific region, is likely to yield more favorable results.

Limitations

To the best of our knowledge, this is the first study based on the most recent GSHS data (2009鈥�2019) and using the same standardized sampling methods across 59 countries that assessed the association of tobacco and/or alcohol use with total and risky sexual behaviors among adolescents on a global scale. We compared the associations according to age, sex, income, and WHO region to yield more precise empirical data for informing policy decisions. However, several limitations should also be considered. First, the cross-sectional design of this study impedes inference of a causal association between tobacco and/or alcohol use and sexual behaviors. Future prospective studies are needed to validate this association. Second, data on tobacco use, alcohol use, and sexual behaviors were collected using self-reported questionnaires, which are susceptible to recall bias and social desirability. Future studies should consider validated biomarkers such as nicotine for tobacco use, phosphatidylethanol for alcohol use, and vaginal swabs of prostate-specific antigens for unprotected sexual behaviors. Third, some unobserved or unmeasured potential confounding factors, such as knowledge about HIV and normative beliefs regarding condom use [50], may influence the association between tobacco and/or alcohol use and risky sexual behaviors. These factors should be taken into consideration in future studies. Fourth, because the GSHS did not provide data on adolescents who dropped out of school, we only included those who were in school, which limits the generalizability of our results. Fifth, due to the limited information available in GSHS data, we were unable to examine which specific types of tobacco and alcohol use among adolescents contributed most to sexual behaviors, thereby warranting further investigation.

Conclusions

Our study highlights a positive association between frequent tobacco and/or alcohol use and total and risky sexual behaviors among adolescents worldwide. Specifically, the strength of these associations varied according to age, sex, and WHO regions. Therefore, health intervention programs, government policies, and advocacy efforts aimed at preventing risky sexual behaviors should focus on reducing tobacco and/or alcohol use among adolescents worldwide, with particular attention to adolescent girls who use tobacco, younger adolescents who use alcohol, and adolescents in the Western Pacific region who use tobacco and/or alcohol.