Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 1.
Published in final edited form as: Drug Alcohol Depend. 2020 Oct 26;218:108383. doi: 10.1016/j.drugalcdep.2020.108383

Latency to cannabis dependence mediates the relationship between age at cannabis use initiation and cannabis use outcomes during treatment in men but not women.

Brian J Sherman 1, Nathaniel L Baker 2, Katherine Schmarder 1, Aimee L McRae-Clark 1, Kevin M Gray 1
PMCID: PMC7750256  NIHMSID: NIHMS1641356  PMID: 33183908

Abstract

Background:

Time from first cannabis use to cannabis dependence (latency) may be an important prognostic indicator of cannabis-related problems and treatment outcomes. Gender differences in latency have been found; however, research in this general area is limited. As cannabis use increases and perceived risk declines, a better understanding of how these factors interact in predicting treatment outcomes is critical.

Methods:

A secondary data analysis of a randomized, double-blind, placebo-controlled pharmacotherapy trial for cannabis dependence (N=302) examined the associations between age of cannabis use onset, time to cannabis dependence (latency), and gender on cannabis use during the trial. Mediation analysis tested whether the association between age of onset and cannabis use during the trial was mediated by latency to cannabis dependence differentially for men and women.

Results:

Age of use initiation was inversely correlated with latency to dependence prior to treatment [HR(95% CI)=1.18 (1.06, 1.30); p=.002] and cannabis use during treatment (β=−1.27; SE=0.37; p<.001). There was a significant mediation effect between age of onset, latency, and cannabis use that varied by gender. Earlier age of onset predicted longer latency, and subsequently, greater cannabis use during the trial in men (21.4% mediated; p<.05), but not women. Other substance use, race, and past psychiatric diagnosis did not predict latency either independently or in interaction models.

Conclusion:

Findings support existing evidence that early cannabis use onset is associated with worse outcomes and add new knowledge on the differential associations between age of onset, latency to cannabis dependence, and treatment outcomes for men and women.

Keywords: Cannabis, cannabis use disorder, dependence, telescoping, age of onset, latency, marijuana, treatment

1. Introduction

Cannabis use in the United States has doubled from 2001–2002 to 2012–2013, with a corresponding increase in prevalence rates of cannabis use disorder (CUD) from 1.5% to 2.9% over that time (Hasin et al., 2015). Myriad physical and psychiatric consequences have been identified (Volkow et al., 2014), including disruption of the endocannabinoid system, which plays a critical role in neurodevelopment and cognition (Curran et al., 2016; Sagar & Gruber, 2018). Time from cannabis initiation to CUD (latency) is faster than for tobacco and alcohol, and comparable to that of cocaine (Lopez-Quintero et al., 2011). Latency to CUD may be an important prognostic indicator whereby more rapid progression increases the risk of drug-related problems and treatment resistance, as has been shown with respect to alcohol (Kay et al., 2010) and heroin (Stoltman et al., 2015). As cannabis use increases, understanding factors associated with latency to CUD can provide a clearer picture of use trajectories and clinical outcomes.

Age at cannabis use onset is a cardinal variable in determining cannabis use trajectories, consequences of use, and potentially, treatment outcomes (Butterworth, Slade & Degenhardt, 2014; Richmond-Rakerd, Slutske & Wood, 2017). Kosty, Seeley, Farmer, Stevens & Lewinsohn (2017) found that the onset and progression of CUD from ages 14–30 display three distinctive trajectories; 1) persistent increasing risk, 2) maturing out, and 3) stable low risk. The persistent increasing risk trajectory was associated with later age of CUD onset as compared to the maturing out trajectory. However, the participants that were categorized into the earlier onset, maturing out trajectory displayed less favorable long-term outcomes, such as lower levels of education and higher uemployment rates. Another study provides evidence that early cannabis use initiation increases the risk for CUD, and that initiation of cannabis use during mid adolescence is associated with long-term, continued use through an individual’s mid thirties (Lee et al., 2017). An epidemiological study conducted in Australia found that earlier age of initiation was associated with an increased risk for development of CUD (Butterworth et al., 2014). Although it was hypothesized that the recent increases in cannabis potency over time would result in shorter latency to CUD, the older birth cohorts, that were presumed to have used lower potency cannabis, demonstrated no significant difference in progression from first use to CUD when compared to the younger cohorts (Butterworth et al., 2014). Another study reported earlier age of cannabis use onset was associated with shorter latency to cannabis dependence (Ehlers et al., 2010); however, that sample was drawn from a family study of alcoholism, which introduces a number of limitations including sampling bias and comorbid alcohol dependence. Given these mixed results and sampling limitations, further investigation and consideration of additional factors that may impact cannabis use onset, trajectory to use disorder, and treatment outcomes is warranted.

One such factor that has been understudied is the impact of sex and gender. Gender differences in cannabis use profiles and latency to CUD are notable (Khan et al., 2013; Sherman et al., 2017). Gender telescoping – whereby women progress from first use to disorder more rapidly than men – has been found in two nationally representative samples in the United States (Khan et al., 2013; Wagner & Anthony, 2007). Likewise, one study found that women progressed from cannabis initiation to treatment entry faster than men (Hernandez-Avlia, et al., 2004). However, a nationally representative sample in Australia found no evidence of gender telescoping (Butterworth et al., 2014). The National Epidemiological Survey on Alcohol and Related Conditions (NESARC) also examined comorbid CUD and collected data on potential covariates in a large sample (N = 11,272). The survey found that the lifetime cumulative probability of progressing to CUD was 27 percent, and that males were at a significantly greater risk of progressing to CUD when compared to female users (Feingold, Livne, Juürgen & Shaul, 2020). A study examining alcohol use trajectories and CUD suggested that the increased risk for CUD in males can be explained by gender roles and socialization, where men are expected to be risk-takers as a part of securing their gender identities (Lee, Brook, De La Rosa, Kim & Brook, 2017). Studies also suggest gender differences in treatment outcomes. Women demonstrated worse cannabis use outcomes than men in a study of buspirone for cannabis dependence (McRae-Clark et al., 2015), in spite of greater self-reported motivation to change (Sherman et al., 2016).

Together, evidence suggests that age of onset is associated with cannabis use trajectories and psychosocial outcomes, and that gender is associated with differential latency to CUD and treatment efficacy. However, studies examining how these factors interact to predict clinical outcomes are lacking, and given the mixed results and limitations of the existing evidence, further research into these associations is critical. The current study examined the role of age of cannabis initiation and gender on latency to cannabis dependence, and the effect of latency on weekly cannabis use in a sample of treatment-seeking adults (N=302) participating in a double-blind, placebo-controlled, pharmacotherapy trial for cannabis dependence (Gray et al., 2017). While evidence infers a causal link between faster latency to CUD (i.e. telescoping) and worse clinical outcomes, to our knowledge this is the first study to examine the effect of latency to cannabis dependence on observed clinical outcomes. Given the evidence discussed above, we hypothesized that 1) female gender and earlier age of initiation would be associated with shorter latency to dependence, 2) shorter latency to dependence would be associated with greater weekly cannabis use during the trial, and 3) latency to dependence would have a mediating effect on the association between age of intitiation and weekly cannabis use during treatment.

2. Methods

2.1. Study Design

The current study analyzed data from the Achieving Cannabis Cessation – Evaluating N-acetylcysteine Treatment (ACCENT) study, a 12-week randomized clinical trial for adults with cannabis dependence (Gray et al., 2017). Participants were randomized to either 1200mg of N-acetylcysteine (NAC) or placebo, and all received medication management and contingency management for urine toxicology confirmed cannabis abstinence. For a full description of the study design see Gray, McClure and colleagues (2014; 2017).

2.2. Participants

Participants were 302 treatment-seeking adults with DSM-IV cannabis dependence recruited from six NIDA CTN locations nationwide. Participants were excluded for 1) allergy or intolerance to NAC, 2) pregnancy or lactation, 3) use of NAC or NAC-containing supplements, 4) use of hazardous concurrent medications, 5) currently enrolled in treatment for cannabis dependence, 6) use of synthetic cannabinoids, 7) other substance dependence, 8) positive urine drug screen (UDS) other than cannabis at randomization (with the exception of amphetamines if the participant had a valid prescription), 9) buprenorphine or methadone maintenance, 10) recent history of asthma, 11) uncontrolled medical or psychiatric illness that could put participant at risk, and 12) current risk of homicide or suicide.

2.3. Assessments

Substance use disorders were assessed using the Diagnostic and Statistical Manual of Mental Disorders – IV checklist (DSM-IV; First, Gibbons, & Williams, 1994) modified for the DSM-5. Lifetime and current psychiatric disorders were assessed using the Mini International Neuropsychiatric Interview Plus (M.I.N.I 6.0; Sheehan et al., 1998). Demographic variables were assessed using standard demographic questions pertaining to age, race, educational level, and other relevant characteristics.

Cannabis Outcomes.

The Timeline Follow-Back (TLFB; Sobell & Sobell, 1992) was used to assess self-reported frequency and quantity of past 30 day marijuana use prior to study initiation and daily use throughout the trial. Mariani and colleagues (2011) procedure was used at the initial screening visit to improve estimation of typical marijuana quantity by providing a gram estimation for each method of marijauana administration used in the past 30 days. Participants then reported on the number of joints, blunts, pipes, vaporizers, spliffs, or other methods used per day.Total grams were estimated by multiplying grams per method of administration, and then summing the daily total across methodsIf participants shared a joint/blunt/etc., partial numbers were reported. Urine cannabinoid tests (UCT) were administered weekly and used to assess biologically-confirmed abstinence.

2.4. Latency to Dependence

Time from cannabis use initiation to cannabis dependence diagnosis was operationalized as “latency to dependence” and calculated by subtracting participant self-reported age at first use from age when DSM-IV criteria for cannabis dependence were first met; cannabis dependence occurred prior to study entry and data collection.

2.5. Data Analytic Plan

Age at cannabis use initiation was used to determine whether earlier age of onset was associated with a more rapid progression to cannabis dependence. Additionally, it was of interest to assess the association of reported latency with cannabis use characteristics and clinical data collected at study entry (following progression to dependence) as well as the effects of latency on cannabis use outcomes during study treatment. Demographic, clinical, and use characteristics were tabulated for each participant at study screening, including age at cannabis use initiation and age at which dependence was noted (latency to dependence). During study treatment, cannabis use was determined weekly as average daily use in grams and negative urine drug screens (<50 ng/ml). Out of the 302 participants included in this analysis, there were 277 with at least one study visit available for study analysis (92%). The average number of attended study visits in this cohort was 10 (SD=3; Range 1– 12) and 61% (170) attended all 12 treatment visits.

2.5.1. Age at Cannabis Use Initiation and Demographic Characteristics

Prior to the primary latency analysis, univariate models were developed to assess the relationship between characteristics measured at study entry and age at cannabis use initiation. Characteristics were stratified by the age at use initiation with tertiles (ties in the lower group: early age at use initiation, ≤14 years; moderate, 15–16 years; and late age, ≥17 years). A two sample Kruskal-Wallis test was used to compare age groups with subsequently measured continuous and ordinal characteristics while a Pearson’s Chi-Square test was used to compare categorical characteristics. Further, age at cannabis use initiation was assessed for a prospective association with weekly measures of cannabis use outcomes during study treatment (Grams used, negative UCT). To assess if age at use initiation was associated with grams used between weekly study visits, generalized linear mixed effects models were developed. To assess the relationship between age at use initiation and urine drug screen results, a repeated-measures logistic regression model was used to analyze the odds of a negative UCT as an indicator of abstinence across all 12 weeks of treatment. At each week, the primary outcome was an indicator of whether the UCT at that visit was negative (<50 ng/mL). Initial model covariates include randomized treatment assignment, week of study visit and the number of days since last visit contact. Additional candidate predictors were chosen as those that were associated with each cannabis use outcome. Candidate predictors were included in the initial models and those highly correlated with other predictors (collinearity) or non-significant in full model were removed for parsimony.

2.5.2. Age at Cannabis Use Initiation and Latency to CUD

Latency from age at first cannabis use to age at dependence was determined from retrospective baseline cannabis use summaries. The association between age at use initiation with latency to dependence was completed using Cox Proportional hazard regression. Modifying effects of demographic characteristics (gender, race, other substance use, and past psychiatric disorder) on the effect of age at use initiation were tested using appropriate interactions. Hazard ratios for the change in age at cannabis use initiation are shown for a one standard deviation change (1 SD=3.4 years). Possible misspecification of the functional form of age at initiation was tested using cumulative martingale residuals and the model proportional hazards assumption was tested using log-log plots and time dependent interactions. The modifying effect of participant gender on this relationship was assessed using model interactions.

2.5.3. Latency to CUD as a Mediator

A mediation analysis was performed to test the extent that latency to dependence would mediate the association between age at use initiation and cannabis use during study treatment and if this effect would vary by gender. Prior to analysis, all model variables were standardized to a standard normal distribution. The extent of mediation is quantified by the difference in the unadjusted association and the indirect effect. To test the statistical significance of the mediated effect, a bootstrap procedure using 1000 replicates was developed. The mean direct and indirect effects as well as the corresponding 95% confidence intervals were derived from the bootstrap results. Mediation models were assessed for the whole cohort as well as stratified by gender. All statistical analyses were performed using SAS, version 9.4 (SAS Institute, Cary, N.C). Significance for all comparisons was set at a 2-sided alpha level of 0.05.

3.0. Results

3.1. Demographic and Clinical Characteristics

Demographic and clinical variables taken at study screening were assessed for the entire treatment cohort. Participants were M (SD) 30.8 (9.0) years of age, 72% male, and 64% Caucasian. Twenty-four percent had at least one past psychiatric diagnosis, and mean age at first cannabis use was 15.2 (3.4) years with an average latency from first use to dependence of 5.6 (6.0) years. Of the 302 randomized participants, 2 participants had latency data removed from the analysis as the reported age at dependence occurred prior to the reported age at first use. For descriptive purposes, age at cannabis use initiation was initially categorized by tertile groups (ties assigned to the lower groups) and associations with demographics, clinical and use characteristics are shown in Table 1. Those with early (≤14 years) and moderate (15–16 years) age at use initiation were significantly younger at the study screening visit (p=.044) than those that were older (≥17 years). Participants with younger age at cannabis use initiation were more likely to have younger age at initiation of cannabis dependence (p<.001) but longer latency between use initiation and cannabis dependence (p<.001). Additionally, participants with early age of use initiation reported greater cannabis use amounts at study baseline as compared to those with older use initiation ages (p<.001). Baseline measures of cannabis use (use days, amount per day) were positively associated with cannabis use amounts and negative UCTs during study treatment (p<.05). Further, baseline measures of depression (HAD-D) and craving (MCQ) were positively associated with cannabis outcomes during study treatment (p<.05).

Table 1.

Demographics and clinical characteristics by age at cannabis use initiation.

Age at Use Initiation
Characteristics Overall (n=302) Early ≤14 Yrs. (n=132) Moderate 15–16 Yrs. (n=97) Late ≥17 Yrs. (n=73) P-value
Age at Screening (years) M(SD) 30.8 (9.0) 30.0 (9.0) 30.4 (9.2) 32.8 (8.7) 0.044
Male %(N) 71.5 (216) 70.5 (93) 73.2 (71) 71.2 (52) 0.900
Caucasian %(N) 63.6 (192) 62.9 (83) 60.8 (59) 68.5 (50) 0.575
Any Past Psychiatric Disorder %(N) 23.8 (72) 20.5 (27) 24.7 (24) 28.8 (21) 0.396
Number of Treatment Visits Attended M(SD) 9.3 (4.3) 9.0 (4.6) 8.9 (4.3) 10.2 (3.5) 0.139
Use Summary**
Age at Cannabis Use Initiation (years) M(SD) 15.2 (3.4) 12.6 (1.5) 15.5 (0.5) 19.6 (3.5) --
Age at Dependence (years) M(SD) 20.8 (6.3) 19.3 (6.3) 20.5 (5.6) 23.8 (6.3) <0.001
Latency to Dependence (years) M(SD) 5.6 (6.0) 6.7 (6.4) 5.1 (5.6) 4.3 (5.3) <0.001
Pre-Screen Days of Use (past 30 days) M(SD) 24.9 (6.6) 25.6 (6.2) 24.8 (6.7) 23.7 (7.0) 0.106
Pre-Screen Grams Used per Day (past 30 days) M(SD) 2.6 (5.1) 3.4 (6.7) 2.2 (3.6) 1.8 (2.6) <0.001
Other Substance Use %(N) 25.2 (76) 21.2 (28) 27.8 (27) 28.8 (21) 0.374
Open in a new tab
**

Cannabis use characteristics are measured as study entry, subsequent to the progression to cannabis dependence.

3.2. Age Cannabis Use Initiation, Gender and Latency to CUD

Age at cannabis use initiation was significantly associated with the latency to dependence [HR (95% CI) = 1.16 (1.05–1.28); p=.003] in the univariate hazard model. Following adjustment for previous psychiatric comorbidities, gender, and the use of other substances, the relationship between age at use initiation and latency to dependence remained significant [HR (95% CI) = 1.18 (1.06–1.30); p=.002]. When stratified by gender, the relationship between age at use initiation and latency to dependence was more robust in men [Adjusted HR (95% CI) = 1.48 (1.27–1.74); p<.001] than women [HR (95% CI) = 1.08 (0.92–1.27); p=.33] (Figure 1). Model interactions between age at use initiation and demographics (race, other substance use and past psychiatric disorder) were found to be statistically insignificant and removed from the final model (all p>0.25). Variable forms met specification criteria (p=.658) and there was no significant evidence that the proportional hazard assumption was violated in the final model (p=.119).

Figure 1.

Figure 1.

Open in a new tab

Years to cannabis dependence by age group at cannabis use initiation stratified by gender.

Data shows the mean and associated standard error of the years from first cannabis use to cannabis dependence by age groups at first use and gender.

*Those in the youngest Male age group [age 8–14, n=93, years to dependence = 6.9 (0.6)] took longer to progress to cannabis dependence than those in the older Male groups; [age 15–16, n=71, years to dependence=4.2 (0.7); p=0.004 and age 17+, n=52, years to dependence=4.5 (0.8); p=0.022].

** Those in the oldest Female age group [17+, n=21, years to dependence=3.9 (1.3)] took less time to progress to cannabis dependence than those in the middle age group [age 15–16, n=26, years to dependence=7.7 (1.1); p=0.028] but not the youngest age group [age 15–16, n=39, years to dependence=6.3 (0.9); p=0.135].

ǂ Females in the middle age group [age 15–16, n=26, years to dependence=7.7 (1.1)] took longer to progress to cannabis dependence as compared to Males [age 15–16, n=71, years to dependence=4.2 (0.7); p=0.009].

Further, increasing age at use initiation was also significantly associated with decreasing weekly use amounts during study treatment (β=−1.25; SE=0.37; p=.001; models adjusted for visit, treatment, baseline cannabis use rates, study treatment, gender and HAD-D) but not increased probability of negative weekly UCT (RR=1.16; 95% CI=0.99, 1.37; p=.07; models adjusted for visit, treatment, baseline cannabis use rates, study treatment and gender). When stratified by gender, the relationship between age at use initiation and weekly use amounts was more robust in men (Adjusted β=−1.55; SE=0.54; p=.005) than women (β=−0.94; SE=0.43; p=.034). The probability of negative weekly UCT with increasing age at use initiation was insignificant for both men (RR=1.22; 95% CI=0.94, 1.57; p=.14) and women (RR=1.13; 95% CI=0.92, 1.40; p=.24). Additionally, increasing latency to dependence was associated with increasing weekly cannabis use amounts during the study β=0.67; SE=0.25; p=.007) and the relationship is significantly modified by gender (Figure 2: p=0.025). Results show that the relationship is driven by increases in use associated with increases in latency in men (β=1.03; SE=0.31; p=.001) but not in women (β=−0.27; SE=0.39; p=.49).

Figure 2.

Figure 2.

Open in a new tab

Cannabis Use Rates during Study Treatment by Latency to Dependence Stratified by Gender

Data shows the mean and associated standard error of the cannabis use by latency to dependence groups stratified by gender. MEN: Those in the longest latency group [long, n=88, Use per Week = 7.5 (0.9) grams] use significantly more cannabis on a weekly basis than those in the moderate and short latency groups; [3–6 years, n=105, use per week=4.5 (0.8); p=.01 and <3 years n=104, use per week=2.6 (0.8); p=.001]. WOMEN: There were no statistical differences in average weekly cannabis use between the 3 latency to dependence tertiles (Short: 4.9 (1.0) vs. Moderate: 3.4 (1.0) vs. Long: 4.2 (0.9); ps>0.28). * p<0.05 compared to Long latency group ǂ p<0.05 compared to men in the same category.

3.3. Latency to CUD as a mediator

To assess the mediating effect of latency to dependence on the relationship between age at use initiation and cannabis use quantities during study treatment, bootstrapped mediation models were assessed. All mediation models are adjusted for study visit, treatment, baseline cannabis use rates, study treatment and gender. The standardized total effect of increasing age at use initiation on weekly use rates during treatment was significant and decreasing in the unadjusted model (β=−0.15; 95% CI=−0.25, −0.05; p<.05) and following inclusion of the mediator, the direct effect was slightly attenuated (β=−0.13; 95% CI=−0.23, −0.05; p<.05) indicating an insignificant mediating effect of latency to dependence on the association (β=−0.015; 95% CI=−0.05, 0.01; p>.05). When stratified by gender, there was a significant total effect in the unadjusted model in men (β=−0.18; 95% CI=−0.32, −0.05; p<.05). Following inclusion of the mediator, the direct effect was attenuated (β=−0.15; 95% CI=−0.27, −0.01; p<.05) indicating an significant, yet partial mediating effect of latency on the association (β=−0.03; 95% CI=−0.01, −0.001; p<.05; 21.4% mediated). This was in contrast to the result in the stratified analysis of women where there was no significant total effect in the model in women (β=−0.10; 95% CI=−0.26, 0.002; p>.05). Following inclusion of the mediator, the direct effect was not statistically changed (β=−0.11; 95% CI=−0.28, −0.001; p<.05) indicating an that no mediating effect of latency on the association was present (β=0.007; 95% CI=−0.009, −0.04; p>.05 ; insignificant mediation). When verifying these results were consistent with baseline reported use amounts (30 day TLFB), results were similar with men having a significant total effect (β=−0.11; 95% CI=−0.20, −0.02; p<.05) whereas women did not (β=−0.07; 95% CI=−0.16, 0.001; p>.05). The indirect mediating effect of latency to dependence was not significant in either case (men β=−0.007; 95% CI=−0.04, 0.02; p>.05; women β=0.004; 95% CI=−0.01, 0.03; p>.05).

4. Discussion

The current study examined the association between age of first cannabis use, latency to cannabis use disorder, and gender on cannabis use outcomes in adults with CUD participating in a 12-week clinical trial of N-Acetylcysteine. Earlier age of initiation was associated with longer latency to dependence and heavier cannabis use upon treatment entry, while later age of initiation was associated with decreased weekly use during treatment. Age of initiation was not associated with biologically confirmed abstinence during the trial. Important and unexpected gender differences were found; longer latency was associated with greater weekly cannabis use in men, but latency was not associated with weekly use in women. Furthermore, latency to CUD mediated the association between age of onset and weekly use during treatment, for men but not women; in men, earlier age of onset was associated with longer latency, which was associated with greater weekly use. Other substance use, race, and past psychiatric diagnosis did not predict latency either independently or in interaction models.

Several important implications are worth noting. First, participants who initiated cannabis use at a younger age took longer to meet criteria for CUD, but still met CUD criteria at a younger age as compared to the older cannabis initiation groups. These participants displayed heavier weekly use during the study, which is consistent with findings displayed in the body of CUD literature that younger age of onset leads to heavier cannabis use (Kosty et al., 2017, Lee et al., 2017). These findings suggest that although rapid progression (i.e. telescoping) has been associated with poor clinical outcomes in alcohol (Kay et al., 2010) and opiate (Stoltman et al., 2015) dependent populations, longer latency may actually confer greater risk in cannabis users. Longer latency from initiation to disorder, combined with heavier use, suggests greater net cannabis exposure, which may exacerbate the deleterious effects of cannabis on the brain. The endocannabinoid system plays a critical role in neurodevelopment and growing evidence points to early, heavy, and prolonged use as significant predictors of subsequent cognitive impairment across domains (see Crane et al., 2013 for review; Becker, Collins, Schultz, Urošević, Schalling & Luciana, 2017). Notably, cognitive impairment is also a risk factor for CUD so the nature of this relationship may be bidirectional. Nevertheless, cannabis-related cognitive deficits due to long-term exposure likely contributes to worse treatment outcomes.

Critically, these associations may be more relevant for males than females as the association between age of cannabis use initiation and cannabis use outcomes was mediated by latency to CUD in men, but not women. Previous research suggests gender telescoping from cannabis initiation to CUD in women (Khan et al., 2013; Wagner & Anthony, 2007). Our findings offer a more nuanced perspective and suggest that when considering gender and rate of progression to CUD, it is critical to also consider age of onset. It is unclear exactly why latency to CUD demonstrated mediation in men but not women, but gender differences in neurodevelopment and the endocannabinoid system may be implicated. For example, neurodevelopment tends to occur earlier in women compared to men (Curran et al., 2016), while CB1 receptor density is higher in men compared to women (Viveros et al., 2012). Furthermore, ovarian hormones have been shown to engender protective effects in human laboratory studies of drug reward (Moran-Santa Maria et al., 2014). This suggests that men may be more susceptible to reinforcing effects and development of tolerance, and therefore may continue to use in larger amounts (resulting in greater net exposure). In contrast, women may be less susceptible to cannabis’ reinforcing effects due to lower CB1 receptor density and the protective effects ovarian hormones, so even if they experiment early in life they may use in smaller amounts and less frequently thus limiting overall exposure.

A final consideration is that latency from first use to cannabis dependence may reflect diverse clinical profiles, rather than vulnerability to addiction per se. Individuals who begin at an early age (8–14 in this study) may have limited access and while they may not progress as rapidly to disorder, they are still being exposed to cannabis’ adverse effects during a critical period of neurodevelopment. In contrast, those who begin at college-age (>17) do so in the context of increased accessibility and peer pressure, and therefore may progress more quickly. Late-onset individuals were also marginally more likely to use other substances which is consistent with the profile of late adolescent-early adult onset heavy user. Lastly, perceived risk of cannabis has decreased notably in recent years and is associated with increased use (Miech, Johnston, & O’Malley, 2017). Younger individuals may minimize the negative consequences of their use, delaying a diagnosis of CUD and treatment-seeking behaviors.

This study has limitations worth noting. The study design relied on retrospective data introducing the possibility of recall bias, the extent of which may itself vary by participant age. Likewise, detailed characteristics of cannabis use history were not ascertained and therefore we cannot account for onset of heavy use, overall net exposure, or significant periods of abstinence. Generalizability is limited to treatment seeking individuals and not the general population. Similarly, participants with serious psychiatric comorbidity were excluded, limiting our ability to examine individuals with serious co-occurring disorders. In addition, a more even subsample distribution across latency groups may have strengthened our findings. A further limitation is that there are challenges with the interpretation of currently available biomarkers of cannabis abstinence (such as UCTs), and time to negative UCT may be impacted by interindividual factors such as metabolism, body mass, and hydration (Loflin et al., 2020). Lastly, the sample was two-thirds male so gender effects should be interpreted with caution. In spite of these limitations, our findings shed new light on the relationship between age of onset, latency to CUD, and cannabis use outcomes.

To our knowledge this is the first study to investigate the associations between age of onset, gender, and latency to CUD on cannabis use outcomes. Findings provide additional support for the risks associated with early age of onset, and critically extend this evidence base by demonstrating gender-moderated mediation whereby early cannabis initiation is associated with longer latency to CUD and greater weekly cannabis use during treatment in men, but not women. Future studies are needed to corroborate these results. Investigation of biological factors that pre-date cannabis onset (e.g. genetic and environmental factors), as well as the nature of the associations between neurodevelopment, early cannabis exposure, and the endocannabinoid system are also warranted. A better understanding of these associations will critically inform cannabis abuse prevention and treatment development.

Highlights.

  • Effects of age of onset, latency to CUD, and gender on cannabis use were examined

  • Earlier onset was associated with longer latency to CUD and heavier use during the trial

  • Latency to CUD mediated the association between age of onset and cannabis outcomes

  • Earlier onset predicted longer latency to CUD, which predicted worse outcomes

  • Longer latency was associated with poor treatment response in men, but not women

Role of Funding Sources

Funding for this study was provided by NIDA Grants U10DA013727 (KMG), K23DA045099 (BJS), and K24DA038240 (ALM). NIDA had no role in the study design, collection, analysis, or interpretation of data, writing the manuscript, and the decision to submit the manuscript for publication.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of Interest

All authors declare they have no conflict of interest.

References

  1. Becker MP, Collins PF, Schultz A, Urošević S, Schmaling B, & Luciana M (2018). Longitudinal changes in cognition in young adult cannabis users. Journal of clinical and experimental neuropsychology, 40(6), 529–543. doi: 10.1080/13803395.2017.1385729 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brook JS, Lee JY, Brown EN, Finch SJ, & Brook DW (2011). Developmental trajectories of marijuana use from adolescence to adulthood: personality and social role outcomes. Psychological reports, 108(2), 339–357. doi: 10.2466/10.18.PR0.108.2.339-357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Butterworth P, Slade T, & Degenhardt L (2014). Factors associated with the timing and onset of cannabis use and cannabis use disorder: results from the 2007 Australian National Survey of Mental Health and Well-Being. Drug Alcohol Rev, 33(5), 555–564. doi: 10.1111/dar.12183 [DOI] [PubMed] [Google Scholar]
  4. Crane N, Schuster R, Fusar-Poli P, & Gonzalez R (2013). Effects of Cannabis on Neurocognitive Functioning: Recent Advances, Neurodevelopmental Influences, and Sex Differences. Neuropsychology Review, 23(2), 117–137. doi: 10.1007/s11065-012-9222-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Curran HV, Freeman TP, Mokrysz C, Lewis DA, Morgan CJ, & Parsons LH (2016). Keep off the grass? Cannabis, cognition and addiction. Nature Reviews Neuroscience, 17(5), 293–306. [DOI] [PubMed] [Google Scholar]
  6. Ehlers CL, Gizer IR, Vieten C, Gilder DA, Stouffer GM, Lau P, & Wilhelmsen KC (2010). Cannabis dependence in the San Francisco Family Study: age of onset of use, DSM-IV symptoms, withdrawal, and heritability. Addict Behav, 35(2), 102–110. doi: 10.1016/j.addbeh.2009.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Feingold D, Livne O, Rehm Juürgen, & Lev-Ran S (2020). Probability and correlates of transition from cannabis use to DSM-5 cannabis use disorder: results from a large-scale nationally representative study. Drug and Alcohol Review, 39(2), 142–151. doi: 10.1111/dar.13031 [DOI] [PubMed] [Google Scholar]
  8. First MS, Gibbon M, & Williams JBW (1994). Structured clinical interview for DSM-IV-TR Axis I disorders, research version, patient edition. SCIP-I/P. New York: Biometrics Research, New York State Psychiatric Institute. [Google Scholar]
  9. Hasin DS, Saha TD, Kerridge BT, Goldstein RB, Chou SP, Zhang H, … Grant BF (2015). Prevalence of Marijuana Use Disorders in the United States Between 2001–2002 and 2012–2013. JAMA Psychiatry, 72(12), 1235–1242. doi: 10.1001/jamapsychiatry.2015.1858 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hernandez-Avila CA, Rounsaville BJ, & Kranzler HR (2004). Opioid-, cannabis-and alcohol-dependent women show more rapid progression to substance abuse treatment. Drug and alcohol dependence, 74(3), 265–272. [DOI] [PubMed] [Google Scholar]
  11. Kay A, Taylor TE, Barthwell AG, Wichelecki J, & Leopold V (2010). Substance use and women’s health. Journal of Addictive Diseases, 29(2), 139–163. doi: 10.1080/10550881003684640 [DOI] [PubMed] [Google Scholar]
  12. Khan SS, Secades-Villa R, Okuda M, Wang S, Pérez-Fuentes G, Kerridge BT, & Blanco C (2013). Gender differences in cannabis use disorders: Results from the National Epidemiologic Survey of Alcohol and Related Conditions. Drug and Alcohol Dependence, 130(1–3), 101–108. doi: 10.1016/j.drugalcdep.2012.10.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kosty DB, Seeley JR, Farmer RF, Stevens JJ, & Lewinsohn PM (2017). Trajectories of cannabis use disorder: risk factors, clinical characteristics and outcomes. Addiction, 112(2), 279–287. doi: 10.1111/add.13557 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lee JY, Brook JS, De La Rosa M, Kim Y, & Brook DW (2017). The association between alcohol use trajectories from adolescence to adulthood and cannabis use disorder in adulthood: a 22-year longitudinal study. The American journal of drug and alcohol abuse, 43(6), 727–733. doi: 10.1080/00952990.2017.1288734 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lee JY, Brook JS, Finch SJ, & Brook DW (2015). Trajectories of marijuana use from adolescence to adulthood predicting unemployment in the mid 30s. The American journal on addictions, 24(5), 452–459. doi: 10.1111/ajad.12240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Lopez-Quintero C, de los Cobos JP, Hasin DS, Okuda M, Wang S, Grant BF, & Blanco C (2011). 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 and Alcohol Dependence, 115(1–2), 120–130. doi: 10.1016/j.drugalcdep.2010.11.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. McClure EA, Sonne SC, Winhusen T, Carroll KM, Ghitza UE, McRae-Clark AL, … Gray KM (2014). Achieving cannabis cessation -- evaluating N-acetylcysteine treatment (ACCENT): design and implementation of a multi-site, randomized controlled study in the National Institute on Drug Abuse Clinical Trials Network. Contemp Clin Trials, 39(2), 211–223. doi: 10.1016/j.cct.2014.08.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. McRae-Clark AL, Baker NL, Gray KM, Killeen TK, Wagner AM, Brady KT, … Norton J (2015). Buspirone treatment of cannabis dependence: A randomized, placebo-controlled trial. Drug Alcohol Depend, 156, 29–37. doi: 10.1016/j.drugalcdep.2015.08.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Miech R, Johnston L, & O’Malley PM (2017). Prevalence and attitudes regarding marijuana use among adolescents over the past decade. Pediatrics, 140(6), e20170982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Moran-Santa Maria MM, Flanagan J, & Brady K (2014). Ovarian hormones and drug abuse. Current psychiatry reports, 16(11), 511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Richmond-Rakerd LS, Slutske WS, Lynskey MT, Agrawal A, Madden PA, Bucholz KK, Heath AC, Statham DJ, & Martin NG (2016). Age at first use and later substance use disorder: Shared genetic and environmental pathways for nicotine, alcohol, and cannabis. Journal of abnormal psychology, 125(7), 946–959. doi: 10.1037/abn0000191 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Richmond-Rakerd LS, Slutske WS, & Wood PK (2017). Age of initiation and substance use progression: A multivariate latent growth analysis. Psychology of addictive behaviors : journal of the Society of Psychologists in Addictive Behaviors, 31(6), 664–675. doi: 10.1037/adb0000304 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Sagar KA, & Gruber SA (2018). Marijuana matters: reviewing the impact of marijuana on cognition, brain structure and function, & exploring policy implications and barriers to research. International review of psychiatry (Abingdon, England), 30(3), 251–267. 10.1080/09540261.2018.1460334 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, … Dunbar GC (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59(Suppl 20), 22–33. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=1998-03251-004&site=ehost-live [PubMed] [Google Scholar]
  25. Sherman BJ, Baker NL, & McRae-Clark AL (2016). Gender differences in cannabis use disorder treatment: change readiness and taking steps predict worse cannabis outcomes for women. Addictive Behaviors, 60, 197–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sherman BJ, Baker NL, & McRae-Clark AL (2017). Effect of oxytocin pretreatment on cannabis outcomes in a brief motivational intervention. Psychiatry research, 249, 318–320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Stoltman JJ, Woodcock EA, Lister JJ, Greenwald MK, & Lundahl LH (2015). Exploration of the telescoping effect among not-in-treatment, intensive heroin-using research volunteers. Drug Alcohol Depend, 148, 217–220. doi: 10.1016/j.drugalcdep.2015.01.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Substance Abuse and Mental Health Services Administration (2014). Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings. Rockville, MD: Substance Abuse and Mental Health Services Administration. [Google Scholar]
  29. Viveros MP, Llorente R, Suarez J, Llorente-Berzal A, Lopez-Gallardo M, & Rodriguez de Fonseca F (2012). The endocannabinoid system in critical neurodevelopmental periods: sex differences and neuropsychiatric implications. Journal of Psychopharmacology, 26(1), 164–176. [DOI] [PubMed] [Google Scholar]
  30. Volkow ND, Baler RD, Compton WM, & Weiss SR (2014). Adverse health effects of marijuana use. The New England journal of medicine, 370(23), 2219–2227. doi: 10.1056/NEJMra1402309 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Wagner FA, & Anthony JC (2007). Male-female differences in the risk of progression from first use to dependence upon cannabis, cocaine, and alcohol. Drug and Alcohol Dependence, 86(2–3), 191–198. doi: 10.1016/j.drugalcdep.2006.06.003 [DOI] [PubMed] [Google Scholar]

RESOURCES