Psychiatric comorbidities can negatively impact on symptom progression and outcomes in individuals with ADHD. Substance-use disorder is commonly comorbid with ADHD, with a prevalence of 45% in adults with ADHD (Jacob et al. 2007). A diagnosis of ADHD has been shown to significantly increase the likelihood of substance use, abuse and dependence in adolescents and adults, with cannabis the most commonly used illicit drug in this population (Molina et al. 2013). Both ADHD and cannabis use have complex aetiologies, with genetic and environmental components, and it is possible that they share common genetic influences. However, despite evidence of a link between these two traits, to date, no common genetic risk factors have been identified. This study aimed to estimate the genetic correlation between ADHD and cannabis use, identify shared genetic factors between the two traits, and examine the causal role that ADHD appears to play in cannabis use in later life by using Mendelian randomisation.*
Genetic data for both ADHD and cannabis use were obtained from recent, large meta-analyses of genome-wide association studies (GWAS) for each trait: the European ancestry subgroup of the Psychiatric Genomics Consortium and iPSYCH (PGC + iPSYCH; 18,345 cases and 32,454 controls) for ADHD; and the International Cannabis Consortium (ICC; 14,374 cases and 17,956 controls) for lifetime cannabis use. The studies included in the two meta-analyses used different criteria for genetic variant filtering; in the current study, variants with different alleles in PGC + iPSYCH and ICC, those with AT/GC alleles, or those within the human leukocyte antigen region of chromosome 6 were removed, resulting in a total of 5,009,020 genetic markers for consideration.
Linkage disequilibrium (LD) score regression analysis was used to estimate both the single-nucleotide polymorphism (SNP)-based heritability of each of ADHD and lifetime cannabis use, and the genetic correlation between the two traits, using data for 1,064,988 genetic markers. To identify genetic factors shared between the two traits, a cross-trait, fixed-effect, inverse-variance–weighted (IVW) meta-analysis of the results of the PGC + iPSYCH and ICC GWAS meta-analyses was performed, with a random-effects meta-analysis run as a sensitivity analysis. Additionally, a gene-based analysis was run, in which SNPs were assigned to a given gene if they were 10 kb up- or down-stream of that gene, allowing calculation of mean SNP associations per gene, and p values. For the gene-based analysis, a genome-wide significance threshold was set at p = 2.79×10-6, following a Bonferroni correction considering 17,297 genes. A sign test was then performed, which assessed whether the direction of effect of variants known to be associated with ADHD was consistent in the cannabis use analysis, and vice versa. Next, a Mendelian randomisation analysis was conducted to investigate the possible causal relationship between ADHD and cannabis use. The analysis was run in both directions: 1) using ADHD as ‘exposure’ and lifetime cannabis use as ‘outcome’; and 2) using cannabis use as ‘exposure’ and ADHD as ‘outcome’. The average effect on outcome, via exposure, of all exposure-associated genetic variants tested was estimated using IVW analysis. Repeated analyses, removing one variant at a time (leave-one-out analyses), were also performed. The estimates obtained were converted into odds ratios for cannabis use in individuals with ADHD versus those without, assuming an ADHD prevalence of 5%.
The results were as follows:
- The LD score regression analyses provided an SNP-based heritability estimate for ADHD of 26% and for lifetime cannabis use of 9%, and provided evidence for a significant genetic correlation between the two traits (rg = 0.29, standard error = 0.068, p = 1.63 x 10-5), indicating a shared background of common genetic variants.
- The cross-trait, fixed-effects meta-analysis found 16 signals that met the genome-wide significance threshold (p = <5 x 10-8). Of these, there were 9 variants, across 7 regions, which had not met genome-wide significance in either the PGC + iPSYCH or ICC meta-analyses alone. Seven of these variants were located in genetic regions that had previously been implicated as associated with ADHD in the PGC + iPSYCH meta-analysis; however the two remaining significant variants lay in regions not previously implicated by either the ADHD meta-analysis or the cannabis use meta-analysis (rs145108385 in chromosome 5, p = 3.30 x 10-8; and rs4259397 in chromosome 8, p = 4.52 x 10-8).
- In the gene-based analysis, three genes (WDPCP, TMEM161B and ZNF251) met the genome-wide significance threshold in the cross-trait analysis and the sensitivity analysis, but did not in either the ADHD or cannabis use meta-analyses alone. TMEM161B lies in a locus previously identified as associated with ADHD in the single-variant analysis, whereas WDPCP and ZNF251 had not previously been identified as associated with either ADHD or cannabis use. This may suggest a possible role for these genes in the biological mechanisms underlying the association between ADHD and cannabis use.
- The sign test demonstrated that ADHD-associated variants had a consistent direction of effect in cannabis use; however none of the cannabis use-associated variants showed significant results in ADHD, suggesting that cannabis use does not increase the risk of ADHD, but supporting the theory that ADHD is associated with increased risk of cannabis use.
- The Mendelian randomisation analysis showed that there exists a significant causal effect of ADHD upon lifetime cannabis use. Cannabis use was 7.9-times more likely for ADHD versus non-ADHD (95% confidence interval 3.72–15.51), and the leave-one-out analysis demonstrated that this increased likelihood is not driven by a single genetic variant. No evidence for a causal effect of cannabis use upon ADHD was found.
A number of potential limitations should be considered when interpreting the results of this study; firstly, the authors acknowledged that the lack of evidence for a causal effect of lifetime cannabis use on ADHD could be solely down to lack of statistical power, since the sample size of the ICC meta-analysis was smaller than that of the PGC + iPSYCH meta-analysis. Additionally, the limited number of variants included in the Mendelian randomisation imparts uncertainty upon the estimate of causal effect of ADHD on cannabis use, as reflected by the wide confidence interval. Furthermore, although this study demonstrated a causal role of ADHD in cannabis use, no information surrounding specific ADHD symptoms/presentations or comorbid disorders was provided, and given that these factors may influence the risk of substance misuse in individuals with ADHD, their role in this causal effect could be further studied.
The authors concluded that the results of this study demonstrate a genetic correlation between ADHD and lifetime cannabis use, and support a causal effect of ADHD on cannabis use. The authors emphasised that these findings highlight the need to take substance misuse into consideration when diagnosing and treating ADHD, and suggested that further genetic studies should be undertaken to further explore the underlying biological mechanisms shared by both traits.
*Mendelian randomisation utilises genetic variants that are known to be associated with a given exposure (i.e. in this study, ADHD or cannabis use), and can therefore be considered as unconfounded proxies for the exposure, in order to test whether that exposure causes a given outcome (i.e. cannabis use or ADHD); the rationale behind this is that alleles are inherited randomly by offspring, therefore avoiding reverse causation or other potentially confounding issues
Jacob CP, Romanos J, Dempfle A, et al. Co-morbidity of adult attention deficit/hyperactivity disorder with focus on personality traits and related disorders in a tertiary referral center. Eur Arch Psychiatry Clin Neurosci 2007; 257: 309-317.
Molina BS, Hinshaw SP, Eugene Arnold L, et al. Adolescent substance use in the multimodal treatment study of attention-deficit/hyperactivity disorder (ADHD) (MTA) as a function of childhood ADHD, random assignment to childhood treatments, and subsequent medication. J Am Acad Child Adolesc Psychiatry 2013; 52: 250-263.
Soler Artigas M, Sánchez-Mora C, Rovira P, et al. Attention-deficit/hyperactivity disorder and lifetime cannabis use: genetic overlap and causality. Mol Psychiatry 2019; Epub ahead of print.