Age of Cannabis Use Onset and Adult Drug Abuse Symptoms: A Prospective Study of Common Risk Factors and Indirect Effects

Abstract

French

The present study examined 1) whether the associations between cannabis use (CU) age of onset and drug abuse by 28 y remain when controlling for risk factors in childhood, adolescence and early adulthood; and 2) the developmental pathways from early risk factors to drug abuse problems. Participants from a longitudinal sample of boys of low socioeconomic status (N = 1,030) were followed from 6 to 28 y. We examined the self-reported CU onset between the ages of 13 and 17 y and drug abuse symptoms by 28 y. The odds of developing any drug abuse symptoms by 28 y were reduced by 31% for each year of delayed CU onset (OR = 0.69). Cannabis, alcohol and other drug frequency at 17 y mediated this association. Still, even when taking that frequency of use into account, adolescents who started using cannabis before 15 y were at a higher risk of developing drug abuse symptoms by age 28 y. Significant indirect effects were found from early adolescent delinquency and affiliation with deviant friends to drug abuse symptoms at 28 y through CU age of onset and substance use frequency at 17 y. The results suggest more clearly than before that prevention programs should aim at delaying CU onset to prevent or reduce drug abuse in adulthood. Furthermore, prevention programs targeting delinquency and/or affiliation with deviant friends in childhood or early adolescence could indirectly reduce substance abuse in adulthood without addressing substance use specifically.

Keywords

cannabis marijuana early onset substance abuse disorder

Cannabis is the most widely used substance worldwide after alcohol and tobacco, including among Canadian adolescents.^1
–3 Indeed, 10% of Canadian adolescents consumed cannabis in Grade 8, a rate that increases to 29% by Grade 12.³ This early-onset cannabis use (CU) may be problematic, as it is associated with several adverse outcomes, including other drug use and substance use disorders.⁴ However, most studies to date on the association between adolescent CU and later substance use defined early-onset as any CU in adolescence without examining age of onset specifically. Whereas 2 studies suggest that an earlier CU age of onset is associated with higher risk of cannabis abuse and dependence as well as alcohol use problems, differences between earlier and later onset were not formally tested^5,6 and one of these studies relied partly on retrospective data,⁵ which is less reliable than prospective data.⁷ Two more studies using the same sample and comparing CU onset before and after 16 y found that CU onset before 16 y was positively associated with cannabis dependence at 20 y⁸ and CU frequency and dependence at 24 y.⁹ However, using dichotomous variables with a cut-off at 16 y may mask significant information about age of onset across adolescence.

Studies on the role of CU on other outcomes and on alcohol use age of onset highlight the importance of examining age of onset throughout adolescence instead of before and after an arbitrarily chosen cut-off age. Indeed, studies found an earlier CU age of onset during adolescence was associated with age at onset of psychosis,^10,11 decline in verbal IQ and tasks tapping trial and error learning and reward processing, and lower rates of high-school graduation.¹² Studies on alcohol use age of onset also report that an earlier alcohol use age of onset is associated with higher odds of later alcohol and drug dependence.^13
–15 Although these studies suggest that examining age of onset across adolescence may be particularly important to fully understand the association between adolescent substance use and future outcomes, no study so far has formally examined the association between CU age of onset and future substance use disorders. Such information is needed to improve our understanding of the pathogenesis of substance use disorders. Indeed, more complete information on the significance of an earlier onset of CU will better guide prevention and intervention programs.

It is also important to look at explanatory variables and mechanisms linking CU onset with later substance use problems, especially considering that few cannabis users become dependent and develop substance use problems.^16,17 This can help determine whether the association between CU age of onset and later substance use problems can be explained by common early risk factors for both early onset CU and later drug abuse, or other risk factors for substance use disorders.⁴ One study has found that the association between CU onset and cannabis dependence was explained by antisocial behaviours and smoking.⁸ However, other studies found that the association between early CU onset and later substance use was not explained by school grades, antisocial behaviours, parent education,⁶ parental separation, parent smoking,⁹ or alcohol, cigarette, and illicit drug use.^5,6 Although these previous studies collectively included numerous covariates, each study controlled only for a limited number. Furthermore, studies did not include other important adolescent covariates of substance use, such as parental supervision,¹⁸ impulsivity,¹⁹ and affiliation with deviant peers.²⁰ Besides simply entering these variables as covariates, understanding their role in developmental pathways from early risk factors to later substance use could help identify prevention and intervention targets.

The Present Study

This study aims to investigate the associations between CU age of onset and drug abuse symptoms by 28 y by examining: 1) whether the associations between CU age of onset and drug abuse symptoms by 28 y remain when taking into account (a) early common risk factors (6 to 13 y), (b) cannabis and other substance use frequency (17 y), and (c) explanatory variables in early adulthood (20 to 23 y); and 2) the developmental pathways (indirect effects) from early common risk factors to drug abuse symptoms by 28 y through CU age of onset, substance use frequency during adolescence, and explanatory variables in early adulthood.

Method

Participants

Participants were from the Montreal Longitudinal and Experimental Study of boys of low socioeconomic status (MLES). In 1984, 1,037 French-speaking boys (mean age [SD], 6.1 [0.31] y) were recruited from schools in low socioeconomic neighbourhoods of Montreal, Canada (for more information, see Tremblay et al.²¹). Sex, ethnicity (mostly Caucasian boys), and socioeconomic status were homogeneous because of the selection procedure. Sample characteristics are presented in Table 1.

Table 1.

Sample Characteristics and Substance Use Prevalence.

Sample Characteristics^a
Mother age at first child	23.3 (4.1)
Father age at first child	26.4 (5.1)
Mother occupational prestige^b	38.3 (12.1)
Father occupational prestige^b	39.4 (12.8)
Family adversity^c	0.33 (0.25)
Intact family structure, % (n)	68.5 (683)
Substance use prevalence
Any CU during adolescence, % (n)	58.3 (484)
Onset at 13 y	3.7 (31)
Onset at 14 y	8.9 (74)
Onset at 15 y	16.9 (140)
Onset at 16 y	16.4 (136)
Onset at 17 y	12.4 (103)
Cigarette use at 17 y, % (n)	47.6 (376)
Alcohol use at 17 y, % (n)	79.7 (628)
Other drug use at 17 y, % (n)	23.7 (186)
Any drug abuse symptoms by 28 years, % (n)	36.9 (198)
Drug abuse symptoms by age of onset of CU, % (n)
Onset at 13 y	76.9 (10)
Onset at 14 y	65.0 (26)
Onset at 15 y	44.3 (35)
Onset at 16 y	41.4 (29)
Onset at 17 y	47.8 (32)
No CU during adolescence	23.9 (52)

CU, cannabis use

^aData are mean (SD) unless indicated otherwise

^bOccupational prestige based on Canadian norms; scale range 17.8 to 101.7; population average 42.7, SD 13.3.

^cFamily adversity range is 0 -1.

Boys were assessed annually from ages 12 to 17 y, and at ages 20 and 28 y; 952 boys participated at least once during the adolescent period and 536 participated at 28 years. The total sample for this study consists of 1,030 boys with data on any outcome variable between 17 and 28 y; i.e., CU age of onset, high school graduation, substance use at 17 y, delinquency at 20 y and drug abuse symptoms at 28 y. The University of Montreal and CHU Ste-Justine Institutional Review Boards approved this project. Written informed consent from parents and teachers, as well as verbal assent from the boys, were obtained during adolescence. In adulthood, written consent was obtained from all participants.

Measures

Drug abuse symptoms

At 28 y, participants completed a domain-specific scale adapted from the Problem Severity scale of the Personal Experience Screening Questionnaire,²² which measures problems associated with the use of any drug (excluding alcohol and tobacco) over the past 12 mo. The scale includes 7 items referring to the absence (scored 0) and presence (scored 1) of problems over the past 12 mo. These items were used to compute a dichotomous variable representing the presence or absence of drug abuse symptoms corresponding to DSM-IV criteria (0 = abuse symptoms absent; 1 = at least one abuse symptom present).

CU age of onset

Each year from 13 to 17 y, adolescents were asked whether they had consumed cannabis in the last 12 mo. These scores were used to compute an age of onset score, representing the first year the adolescents answered that they consumed cannabis. Computed scores ranged from 1 (CU onset by 13 y) to 6 (no CU onset by 17 y).

Early common risk factors

Paternal alcohol use problems were assessed when the boys were 12 y through an interview with the mother and a subset of fathers. Trained interviewers conducted the interview through a telephone survey using the Short Michigan Alcoholism Screening Test (SMAST).²³ The SMAST allowed mothers to rate fathers reliably, as supported by a study of a subset of 160 fathers showing a high concordance between mother-reported and father-reported alcohol use problems in this sample.²⁴

Family adversity at 6 y was assessed using 6 variables reflecting the quality of the family environment: 1) mothers’ occupational status, 2) fathers’ occupational status, 3) mothers’ educational level (number of years in school), 4) father’s educational level, 5) mothers’ age at the birth of their first child, and 6) family structure (intact or non-intact).

Academic achievement at 12 y was obtained from elementary school teachers, who reported on students’ achievement in French (first language) and mathematics. Scores ranged from 1 (academic failure) to 5 (excellent academic performance).

Verbal IQ at 13 y was assessed using the Sentence Completion Test.²⁵

Impulsivity at 13 y was assessed through self-reports with a short form of the Eysenck Impulsiveness-Venturesomeness-Empathy scale, which included the 5 items (α = 0.69) with the highest factor loadings on the original scale.^26,27

Affiliation with deviant friends at 13 y was obtained by asking boys whether they were part of a group or a gang that carried out reprehensible acts.

Delinquency at 12 y was measured using 17 items from the Self-Reported Delinquency Questionnaire (SRDQ)²⁸ assessing involvement in delinquent behaviours over the last 12 mo by measuring physical violence (5 items, e.g., used a weapon during a fight), theft (7 items, e.g., stole $100 or more), and vandalism (5 items, e.g., vandalised a car). Each question was rated on a 4-point scale (0 = never; 3 = very often). Cronbach’s alpha for this scale was 0.79.

Parental supervision at 12 years was measured using 2 items (α = 0.73) asking ‘Do your parents know where you are when you are not at home?’ and ‘Do your parents know who you are with, when you are not at home?’ rated on a scale ranging from 1 (never) to 4 (always).

Substance use frequency at 17 y

Cigarette, alcohol, cannabis and other drug use frequency at 17 y was obtained by asking adolescents whether they had smoked cigarettes, drunk alcohol, consumed cannabis and consumed drugs other than cannabis (hallucinogens, cocaine, amphetamines, barbiturates, tranquillisers, heroin, inhalants, and other drugs) in the last 12 mo (1 = never, 7 = 40 or more times).

Explanatory variables in early adulthood

Data on high school graduation was obtained when the boys were 23 y old from the Education Department’s official records. Because the probability of completing secondary school dramatically decreases after 20 to 24 y,²⁹ graduation status at 23 y is likely to be definitive for most participants.

Finally, delinquency at 20 y was measured using the Self-Reported Delinquency Questionnaire (see early common risk factors section). Cronbach’s alpha for this scale at 20 y was 0.63.

Data Analyses

Analyses were carried out using Mplus 7.0.³⁰ The estimator used was maximum likelihood with robust standard error estimation (MLR), except for models testing indirect effects, which used weighted least squares means and variance adjusted estimation (WLSMV). First, logistic regressions were conducted to test the association between CU age of onset and drug abuse symptoms, and whether other variables reduced this association. Regressions were conducted in 4 steps: a first model with CU age of onset only, a second model with early common risk factors added to the model, a third model including substance use frequency, and a final model including early adulthood explanatory variables. Path models were conducted to examine indirect effects. Indirect effects were tested in Mplus with a 1,000 samples bias-corrected bootstrap, which reduces bias and yields a more precise type I error.³¹

Results

Descriptive Statistics

Table 1 presents substance use prevalence for the sample. The prevalence of participants with drug abuse symptoms by 28 y decreased with increasing age of CU onset, dropping from 77% in boys who initiated CU at 13 y or before to 48% in boys who initiated CU at 17 y, and to 24% in boys who did not use CU during adolescence. A full correlation table and descriptive statistics for all study variables is provided as Supplementary Material (Table S1). The number of years cannabis was consumed and the rates of other substance use as a function of CU age of onset can also be found in Supplementary Tables S2 and S3.

Does the Association Between CU Age of Onset and Drug Abuse Symptoms by 28 Y Remain When Accounting for Other Risk Factors?

The results from the logistic regressions predicting the presence of any drug abuse symptoms are presented in Table 2. Without considering any covariates, the odds of developing any drug abuse symptoms by 28 y were reduced by 31% for each year of delayed CU onset (model 1; R² = 0.09). Early common risk factors (model 2; ΔR² = 0.02) were not directly associated with drug abuse symptoms by 28 y, and their inclusion did not affect the association between CU age of onset and later drug abuse symptoms. Once substance use frequency at 17 y was added to the model (model 3; ΔR² = 0.09), the contribution of CU age of onset was reduced to non-significance. Cannabis use frequency at 17 y was associated with 1.32-times greater odds of reporting any drug abuse symptoms by 28 y. When early adulthood explanatory variables were added to the model (model 4; ΔR² = 0.03), only alcohol use frequency at 17 y reached significance (OR = 0.87).

Table 2.

Results from Logistic Regressions Predicting Drug Abuse Symptoms by 28 Years.

	Model 1			Model 2			Model 3			Model 4
	B	OR	95%CI	B	OR	95%C I	B	OR	95%CI	B	OR	95%CI
CU age onset	−0.37^c	0.69	0.61–0.78	−0.37^c	0.69	0.60–0.79	−0.06	0.95	0.74 –1.14	−0.04	0.97	0.80–1.17
Father AUP				0.25	1.28	0.69–2.36	0.19	1.21	0.63–2.30	0.19	1.21	0.62–2.33
Family adversity 6 y				0.10	1.10	0.47–2.58	0.05	1.05	0.42–2.60	−0.05	0.95	0.37–2.45
Academic achievement 12 y				−0.11	0.89	0.73–1.09	−0.05	0.95	0.77–1.17	−0.01	0.99	0.80–1.23
Verbal IQ 13 y				0.10	1.10	0.98–1.23	0.10	1.11	0.98–1.25	0.12	1.12	0.99–1.28
Impulsivity 13 y				−0.04	0.96	0.84–1.10	−0.04	0.96	0.84–1.10	−0.04	0.96	0.84–1.11
Deviant friends 13 y				−0.11	0.90	0.52–1.57	−0.12	0.89	0.50–1.57	−0.23	0.80	0.43–1.46
Delinquency 12 y				−0.01	0.99	0.93–1.04	0.00	1.00	0.95–1.06	−0.01	0.99	0.93–1.05
Parental supervision 12 y				−0.08	0.92	0.80–1.10	−0.07	0.93	0.80–1.08	−0.06	0.94	0.81–1.10
Cigarette frequency 17 y							0.06	1.07	0.98–1.16	0.06	1.06	0.97–1.16
Alcohol frequency 17 y							−0.11	0.90	0.79–1.01	−0.14^a	0.87	0.77–0.99
Cannabis frequency 17 y							0.28^c	1.32	1.14–1.53	0.30^c	1.35	1.16–1.59
Other drug frequency 17 y							0.13	1.14	0.94–1.39	0.11	1.12	0.91–1.37
High School graduation										−0.22	0.80	0.47–1.35
Delinquency 20 y										0.13	1.14	0.99–1.31

AUP, alcohol use problems; CU, cannabis use; y, years.

^a P < 0.05, ^b P < 0.01, ^c P < 0.001.

Odds of Having Any Drug Abuse Symptoms by Age of CU Onset

To examine whether a specific age of CU onset was associated with greater odds of having any drug abuse symptoms, the odds of developing any drug abuse symptoms by age 28 were examined by age of CU onset, with each age compared to later CU (see Table 3). The odds of developing any drug abuse symptoms by age 28 y were non-significant if CU had its onset between 15 and 17 y but were significant and almost doubled each year if onset was before 15 y (OR = 1.97, onset by 14; OR = 3.44, onset by 13). Indeed, rates of having any drug abuse symptoms by 28 y were 68% in cannabis users who started early (at or before 14 y) compared to 44% in users who started later (χ² [1, N = 269] = 9.39, P = 0.002).

Table 3.

Odds of Drug Abuse Symptoms by 28 Years for Each Year of Delayed Onset of Cannabis Use^a.

	Drug Abuse Symptoms
Onset by	OR	95% CI
13 y	3.44	1.05–11.22
14 y	1.97	1.08–3.61
15 y	1.10	0.68–1.80
16 y	0.87	0.51–1.49
17 y	1.14	0.64–2.03

^aComparison group for each analysis is later cannabis use. Analyses control for significant covariates from model 4: alcohol use frequency and cannabis use frequency at 17 y.

To clarify how CU frequency contributed to the increased risk of developing any drug abuse symptoms in adolescent cannabis users, a cross-tabulation across different patterns of cannabis use frequency was conducted. Within adolescent cannabis users, rates of having any drug abuse symptoms by 28 y were significantly higher in cannabis users who consumed cannabis frequently at 17 y (20 or more times in the last year) than in cannabis users who consumed less frequently (72% v. 40%; χ² (1, N = 227) = 22.29, P < 0.001).

Is CU Age of Onset Part of an Indirect Pathway from Early Risk Factors to Drug Abuse Symptoms?

When all variables were combined in a path model (see Figure 1), earlier age of CU onset was significantly predicted by higher verbal IQ and delinquency at 12 y, impulsivity and affiliation with deviant peers at 13 y, and having a father with alcohol use problems. In turn, earlier CU age of onset was associated with a lower rate of high school graduation, a higher frequency of cigarette, alcohol, cannabis and other drug use at 17 y, and delinquency at 20 y. Finally, a higher frequency of cannabis and other drug use at 17 y as well as a lower frequency of alcohol use were associated with having any drug abuse symptoms by 28 y.

Figure 1.

Significant direct and indirect paths from early risk factors to drug abuse symptoms by 28 y through cannabis use (CU) age of onset. AUP: alcohol use problems. Double lined arrows indicate indirect paths going through cannabis age of onset. Model fit: χ² = 9.68, dl = 9, P = 0.38; CFI = 1.00; TLI = 0.99; RMSEA = 0.009 (90%CI, 0.000 to 0.037). Correlations among variables within a developmental period (i.e., among early common risk factors, and among substance use frequency variables at 17 y) were included in the model. Standardized coefficients are provided. *P < 0.05, **P < 0.01, ***P < 0.001. Significant indirect effects from CU age of onset to drug abuse symptoms by 28 y included: CU age of onset → CU frequency at 17 y → drug abuse symptoms by 28 y (ab = −0.196, SE = 0.058, P = 0.001); CU age of onset → other drug frequency at 17 y → drug abuse symptoms by 28 y (ab = −0.071, SE = 0.032, P = 0.026); and CU age of onset → alcohol frequency at 17 y → drug abuse symptoms by 28 y (ab = 0.077, SE = 0.034, P = 0.024). Significant indirect effects from early factors through CU age of onset predicting drug abuse symptoms by 28 y included: delinquency at 12 y → CU age of onset → CU frequency at 17 y → drug abuse symptoms by 28 y (abc = 0.035, SE = 0.014, P = 0.011); deviant friends at 13 y → CU age of onset → CU frequency at 17 y → drug abuse symptoms by 28 y (abc = 0.047, SE = 0.016, P = 0.004); deviant friends at 13 y → CU age of onset → other drug frequency at 17 y → drug abuse symptoms by 28 y (abc = 0.017, SE= 0.008, P = 0.038); deviant friends at 13 y → CU age of onset → alcohol frequency at 17 y → drug abuse symptoms by 28 y (abc = −0.018, SE = 0.009, P = 0.039). Other significant direct and indirect paths (predicting high school graduation, other drug frequency, cigarette use frequency, alcohol use frequency and delinquency) are provided in the Supplementary Materials (Figure S1).

Several significant indirect effects were also found (see Figure 1). Significant indirect paths from CU age of onset through CU frequency and other drug use frequency at 17 y positively predicted having any drug abuse symptoms by 28 y, and negatively predicted any drug abuse symptoms through alcohol use frequency at 17 y. Indirect effects from early risk factors to any drug abuse symptoms were also found (see Figure 1). Delinquency at 12 y indirectly predicted having any drug abuse symptoms by 28 y through CU age of onset and cannabis frequency at 17 y. Affiliation with deviant friends at 13 y indirectly predicted having any drug abuse symptoms by 28 y through CU age of onset and cannabis, other drug and alcohol frequency at 17 y.

Discussion

The results of the present study highlight that cannabis users in early adolescence are at higher odds of developing drug abuse symptoms in adulthood. Previous research has shown that any CU during adolescence was associated with more substance use problems later in life⁴ and that onset before 16 y was associated with cannabis dependence.⁹ The present study extends upon those results not only by showing that any CU during adolescence is associated with higher odds of reporting drug abuse symptoms by 28 y, but also that an earlier age of onset increased these odds further. A second contribution of this study was to qualify and quantify the risk conveyed by early age of onset for developing drug abuse symptoms by 28 y by including early risk factors in the models and examining potential explanatory mechanisms. Much of the effect found between CU age of onset and drug abuse symptoms was explained by the frequency of use of cannabis at 17 y, and partly by other drugs. Still, even when taking frequency of use into account, adolescents who started using before 15 y were at higher risk of developing drug abuse symptoms. Furthermore, although earlier CU was associated with higher alcohol use frequency at 17 y, alcohol frequency was negatively associated with drug abuse symptoms by 28 y. Nonetheless, it must be noted that alcohol frequency at 17 y was associated with increased odds of alcohol abuse symptoms by 28 y (OR = 1.16; 95% CI = 1.04 to 1.30; see Appendix Figure S2), suggesting a specific association between alcohol use in adolescence and alcohol problems in adulthood. These specific associations between adolescent alcohol frequency and later alcohol use disorders, but not cannabis or other drug disorders, have previously been shown in other research,^32,33 but future studies are needed to clarify how frequent adolescent cannabis or alcohol use, or both, increase the risk of developing specific substance use problems in early onset substance users (be it early onset-cannabis and/or alcohol users). Finally, we found 2 different independent pathways to drug abuse symptoms through early onset CU: one stemming from high levels of delinquency and the other from an affiliation with deviant peers in early adolescence.

Clinical Implications

Some school-based prevention programs have already been shown to be effective in reducing substance use among 12- to 19-y-old adolescents, the most effective being interactive programs that incorporate a motivational social component, where adolescents notably develop skills to resist peer pressure, but still include other aspects such as knowledge and affective components.³⁴ The finding that a younger CU age of onset during adolescence, particularly between 13 and 15 y, increases the odds of developing drug abuse symptoms stresses the importance of targeting this period for preventing or reducing CU, and thus it may be important to implement these programs by the end of elementary school, at least for the prevention of CU.

However, targeting early adolescence risk factors upstream could also prevent drug abuse symptoms without explicitly addressing substance use. Indeed, because early delinquency and affiliation with deviant friends were indirectly associated with drug abuse symptoms by 28 y through early CU onset, targeting these risk factors before adolescence or in early adolescence, either through community, health, or education approaches, may reduce the odds of initiating CU early and developing later drug abuse symptoms. Whereas studies have already shown that interventions in childhood and early adolescence targeting delinquent behaviours can be effective in reducing substance use during adolescence,^35,36 future studies could look at whether such interventions have longer-term effects on adult substance use problems.

Strengths and Limitations

The main strength of this study was its prospective design, which allowed the examination of the predictive relationships and indirect effects of CU age of onset from a developmental perspective. The use of prospective longitudinal data also allowed for a more reliable measure of CU age of onset, as retrospective reports have been found to be only moderately reliable, especially for participants with an earlier age of onset.⁷ Still, some limitations should be considered. While the current study included various risk factors for substance use and examined indirect effects from early risk factors to drug abuse symptoms in adulthood, there may still be other factors to explain the associations. Notably, since adolescence is an important period for brain development,³⁷ early risk factors and CU may have an impact on neurological development and cognitive function,¹² which could be tested as potential mediators of the association between the risk for CU onset and substance abuse in future studies. Regarding the study design, although its prospective nature was a strength, it remains correlational and thus does not not show causal relationships. The current sample included only French-speaking Caucasian boys from low socioeconomic neighbourhoods in Montreal and the results should be replicated with girls and in more diverse geographical populations. Substance use data were obtained by self-report, which is susceptible to bias, notably social desirability. That said, participants were guaranteed confidentiality and self-reports have been shown as reliable in assessing substance use.³⁸ Furthermore, quantity, potency, and mode of administration of cannabis were not considered and future studies should look at whether these factors affect the association between CU age of onset and drug abuse symptoms. Notably, considering that the potency of cannabis products increased over the last 2 decades³⁹ and that adolescent CU was assessed from 1991 to 1995, it is possible that the higher content of Δ-9-tetrahydrocannabinol in the cannabis available today would be associated with higher rates of drug abuse symptoms.

Conclusions

Previous studies have examined how CU onset before 18 y or 16 y was associated with later substance use problems.^5

–9 The present study builds on this previous data, examining CU onset between 13 and 17 y of age. Our results showed that CU frequency explained much of the association between CU age of onset and later drug abuse symptoms but that CU onset at 13 or 14 y was still associated with higher odds of developing drug abuse symptoms by 28 y, even after controlling for cannabis and other substance use frequency. Explanatory mechanisms were examined and indirect pathways to drug abuse symptoms through early onset CU were found to be associated with delinquency and affiliation with deviant peers in early adolescence. This suggests that targeting these risk factors in late childhood or early adolescence may help prevention efforts aimed at delaying CU.

Supplemental Material

Supplementary_material_final - Age of Cannabis Use Onset and Adult Drug Abuse Symptoms: A Prospective Study of Common Risk Factors and Indirect Effects

Supplementary_material_final for Age of Cannabis Use Onset and Adult Drug Abuse Symptoms: A Prospective Study of Common Risk Factors and Indirect Effects by Charlie Rioux, Natalie Castellanos-Ryan, Sophie Parent, Frank Vitaro, Richard Ernest Tremblay and Jean Richard Séguin in The Canadian Journal of Psychiatry

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was made possible by a scholarship to CR from the Fonds de la Recherche du Québec – Santé and grants from the Canadian Institutes of Health Research (no. MOP 97910), the Social Science and Humanities Research Council of Canada, the National Health Research and Development Program, the Fonds de la Recherche du Québec - Société et Culture, and the Fonds de la Recherche du Québec - Santé. These Funding agencies had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.

Supplemental Material

Supplementary material for this article is available online.

References

Degenhardt

Stockings

Patton

. The increasing global health priority of substance use in young people. Lancet Psychiatry. 2016;3(3):251–264.

Institut de la statistique du Québec. Enquête québécoise sur le tabac, l’alcool, la drogue et le jeu chez les élèves du secondaire, 2013. Sainte-Foy, QC: Publications du Québec; 2014.

Leatherdale

Burkhalter

. The substance use profile of Canadian youth: exploring the prevalence of alcohol, drug and tobacco use by gender and grade. Addict Behav. 2012;37(3):318–322.

Hall

. What has research over the past two decades revealed about the adverse health effects of recreational cannabis use? Addiction. 2015;110(1):19–35.

Behrendt

Wittchen

Hofler

. Transitions from first substance use to substance use disorders in adolescence: Is early onset associated with a rapid escalation? Drug Alcohol Depend. 2009;99(1-3):68–78.

Scholes-Balog

Hemphill

Evans-Whipp

. Developmental trajectories of adolescent cannabis use and their relationship to young adult social and behavioural adjustment: A longitudinal study of Australian youth. Addict Behav. 2016;53:11–18.

Parra

O’Neill

Sher

. Reliability of self-reported age of substance involvement onset. Psychol Addict Beha. 2003;17(3):211–218.

Coffey

Carlin

Lynskey

. Adolescent precursors of cannabis dependence: findings from the Victorian Adolescent Health Cohort Study. Br J Psychiatry. 2003;182:330–336.

Swift

Coffey

Carlin

. Adolescent cannabis users at 24 years: trajectories to regular weekly use and dependence in young adulthood. Addiction. 2008;103(8):1361–1370.

10.

Stefanis

Dragovic

Power

. Age at initiation of cannabis use predicts age at onset of psychosis: the 7- to 8-Year Trend. Schizophr Bull. 2013;39(2):251–254.

11.

Galvez-Buccollini

Proal

Tomaselli

. Association between age at onset of psychosis and age at onset of cannabis use in non-affective psychosis. Schizophr Res. 2012;139(1-3):157–160.

12.

Castellanos-Ryan

Parent

Vitaro

. Adolescent Cannabis Use, Change in Neurocognitive Function, and High-School Graduation: A Longitudinal Study from Early Adolescence to Young Adulthood. Development and Psychopathology. in press.

13.

Grant

Stinson

Harford

. Age at onset of alcohol use and DSM-IV alcohol abuse and dependence: A 12-year follow-up. J Subst Abuse. 2001;13(4):493–504.

14.

King

Chassin

. A prospective study of the effects of age of initiation of alcohol and drug use on young adult substance dependence. J Stud Alcohol Drug. 2007;68(2):256–265.

15.

DeWit

Adlaf

Offord

. Age at first alcohol use: a risk factor for the development of alcohol disorders. Am J Psychiatry. 2000;157(5):745–750.

16.

van der Pol

Liebregts

de Graaf

. Predicting the transition from frequent cannabis use to cannabis dependence: a three-year prospective study. Drug Alcohol Depend. 2013;133(2):352–359.

17.

Lopez-Quintero

de los Cobos

Hasin

. Probability and predictors of transition from first use to dependence on nicotine, alcohol, cannabis, and cocaine: Results of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). Drug Alcohol Depend. 2011;115(1-2):120–130.

18.

Bohnert

Anthony

Breslau

. Parental Monitoring at Age 11 and Subsequent Onset of Cannabis Use Up to Age 17: results from a prospective study. J Stud Alcohol Drug. 2012;73(2):173–177.

19.

Castellanos-Ryan

Conrod

Personality and Substance Misuse: Evidence for a Four-Factor Model of Vulnerability. In: Verster

Brady

Galanter

Conrod

, eds. Drug Abuse and Addiction in Medical Illness: Springer New York; 2012. p 47–62.

20.

Marschall-Lévesque

Castellanos-Ryan

Vitaro

. Moderators of the association between peer and target adolescent substance use. Addict Behav. 2014;39(1):48–70.

21.

Tremblay

Pihl

Vitaro

. Predicting early onset of male antisocial behavior from preschool behavior. Arch General Psychiatry. 1994;51(9):732–739.

22.

Heniy

Winters

. Development of psychosocial scales for the assessment of adolescents involved with alcohol and drugs. Intern J Addict. 1989;24(10):973–1001.

23.

Selzer

Vinokur

van Rooijen

. A self-administered Short Michigan Alcoholism Screening Test (SMAST). J Stud Alcohol. 1975;36(1):117–126.

24.

Vitaro

Dobkin

Carbonneau

. Personal and familial characteristics of resilient sons of male alcoholics. Addiction. 1996;91(8):1161–1177.

25.

Lorge

Thorndike

. The Lorge-Thorndike Intelligence Tests. New York, NY: Houghton Mifflin; 1950.

26.

Eysenck

SBG

Easting

Pearson

. Age norms for impulsiveness, venturesomeness and empathy in children. Personal Individ Diff. 1984;5(3):315–321.

27.

Eysenck

SBG

Eysenck

. Impulsiveness and venturesomeness: their position in a dimensional system of personality description. Psycholo Report. 1978;43(3):1247–1255.

28.

LeBlanc

Fréchette

. Male criminal activity from childhood through youth: Multilevel and developmental perspective. New York, NY: Springer-Verlag; 1989.

29.

Bowlby

McMullen

. Provincial dropout rates - Trends and consequences. Ottawa, ON: Statistics Canada; 2005.

30.

Muthén

. MPlus User’s Guide. Los Angeles, CA: Muthén & Muthén; 1998–2012.

31.

Preacher

Hayes

. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Met. 2008;40(3):879–891.

32.

Kendler

Ohlsson

Edwards

. A developmental model for alcohol use disorders in Swedish men. Psycholo Med. 2016;46(13):2759–2770.

33.

Grant

Scherrer

Lynskey

. Adolescent alcohol use is a risk factor for adult alcohol and drug dependence: evidence from a twin design. Psychol Med. 2006;36(1):109–118.

34.

Porath-Waller

Beasley

Beirness

. A meta-analytic review of school-based prevention for cannabis use. Health Edu Behav. 2010;37(5):709–723.

35.

Conrod

Castellanos-Ryan

Mackie

. Long-term effects of a personality-targeted intervention to reduce alcohol use in adolescents. J Consult Clinical Psychol. 2011;79(3):296–306.

36.

Castellanos-Ryan

Séguin

Vitaro

. Impact of a 2-year multimodal intervention for disruptive 6-year-olds on substance use in adolescence: randomised controlled trial. Brit J Psych. 2013(203):188–195.

37.

Spear

. Adolescent Neurodevelopment. J Adolescent Health. 2013;52(2):S7–S13.

38.

Clark

Winters

. Measuring risks and outcomes in substance use disorders prevention research. J Consult Clinical Psychol. 2002;70(6):1207–1223.

39.

ElSohly

Mehmedic

Foster

. Changes in Cannabis Potency Over the Last 2 Decades (1995-2014): analysis of current data in the United States. Biolog Psych. 2016;79(7):613–619.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB