Over the past 20 years, rates of cannabis use have been increasing among adults in the United States (US). According to the National Survey on Drug Use and Health (NSDUH) (Administration, 2021), 34.5 % of adults age 18–25 reported past year marijuana use in 2020 compared to 29.8 % in 2002. More striking, 16.3 % of adults above age 25 reported past year marijuana use in 2020 compared to 7.0 % in 2002. Likewise, the prevalence of cannabis use disorder (CUD) continues to increase not only among the general US population (Hasin, 2018) but has more than doubled among US military veterans in recent years (Bonn-Miller et al., 2012; Davis et al., 2018; Hill et al., 2021a). Veterans diagnosed with insomnia, anxiety disorders, major depressive disorder (MDD), and posttraumatic stress disorder (PTSD), are particularly inclined to use cannabis as a coping mechanism (Boden et al., 2013; Bonn-Miller et al., 2007; Metrik et al., 2016). In 2020, the rate of lifetime CUD was 9.2 % among veterans, 12.1 % among veterans with PTSD, and 8.9 %−13.1 % in veterans with other psychiatric and substance use disorders (Hill et al., 2021a). Among veterans who report nonmedical cannabis use, the rate of lifetime CUD was even higher (17.4 %) and was associated with several psychiatric and substance use disorders (Browne et al., 2022). This increase is of great concern as CUD has been associated with suicidal behaviors in veterans (Adkisson et al., 2019; Grove et al., 2023; Hill et al., 2021b; Shamabadi et al., 2023), including those from our own cohort (Kimbrel et al., 2017).

Despite the perceived alleviation of trauma-related symptoms, there is a lack of evidence to support the use of therapeutic cannabis for mental health disorders (Black et al., 2019; Solmi et al., 2023). In fact, CUD is associated not only with negative health outcomes such as pulmonary conditions (Aldington et al., 2007; Hancox et al., 2010) and myocardial infarction (Lindsay et al., 2005), but also with lower educational and occupational attainment (Compton et al., 2014; Lynskey and Hall, 2000), as well as a higher likelihood of other substance use disorders (SUDs), anxiety and mood disorders, and PTSD (Gentes et al., 2016; Gunn et al., 2020; Hasin et al., 2016; Livingston et al., 2022; Metrik et al., 2022). Indeed, PTSD is the most highly comorbid psychiatric disorder among veterans with CUD (Bonn-Miller et al., 2012; Cougle et al., 2011). A recent study found the prevalence of PTSD among Iraq/Afghanistan era veterans with a CUD diagnosis was 72.3 % (Bryan et al., 2021). This comorbidity is particularly concerning as PTSD symptom severity has been linked not only to cannabis use, but to severity of cannabis withdrawal and emotion-related cravings (Boden et al., 2013) as well as a slower decline in cannabis use during early cessation attempts (Bonn-Miller et al., 2015).

Understanding the impact of CUD on psychiatric and health outcomes in veterans is of utmost importance. Moreover, the ability to identify testable biomarkers for CUD could greatly improve veteran health management, for both those with PTSD and those without. DNA methylation (DNAm) is an epigenetic modification involving the transfer of a methyl group at the C5 position of a cytosine nucleotide to form 5-methylcytosine and can be measured across the genome at CpG (cytosine-phosphate-guanine) sites. DNAm can be influenced by a variety of factors including genetics and developmental processes, disease and health conditions, and environmental exposures including stress and external toxins. As such, researchers have explored epigenome-wide studies (EWAS) for cannabis and its derivatives in animal models (Murphy et al., 2018; Wanner et al., 2020, 2021; Watson et al., 2015) as well as in human sperm cells (Murphy et al., 2018; Schrott et al., 2021) and whole blood (Clark et al., 2021; Markunas et al., 2021; Nannini et al., 2023; Osborne et al., 2020), but to our knowledge, no EWAS for CUD has been performed in a veteran cohort. The high prevalence of PTSD, CUD, and suicide among veterans, and the complex relationships between these disorders, highlights the importance of investigating DNAm and CUD in this at-risk population.

Here, we examined the association between lifetime CUD and genome-wide DNAm using participants from the Veterans Affairs (VA) Mid-Atlantic Mental Illness Research, Education and Clinical Center (MIRECC) Study of Post-Deployment Mental Health, an Iraq/Afghanistan-era veteran cohort enriched for PTSD. In addition to differential analysis, and owing to the remarkably high co-occurrence of PTSD and CUD that has been reported previously (Bryan et al., 2021), we investigated potential moderating effects of current PTSD status on the relationship between CUD and CUD-associated DNAm probes. Finally, because of the strong correlations between CUD and other psychiatric disorders, we examined potential indirect effects of DNAm on the relationship between CUD and CUD-associated psychiatric disorders, including suicidal thoughts and behaviors.