Most studies on the association of anxiety disorders with substance use are cross-sectional, and a few are longitudinal (Robinson et al. 2011). Epidemiologic studies indicate a two- to threefold increased risk for alcohol use disorders in patients with anxiety disorders. Grant et al. (2005) reported data from the National Epidemiologic Survey on Alcohol and Related Conditions, which studied a large representative sample (N = 43,093) of the adult US population. Prevalence estimates for 12-month and lifetime GAD were 2.1 % and 4.1 %, respectively. Higher rates were found for males. GAD was highly comorbid with substance use in general and with other anxiety, mood, and personality disorders. Specifically, the odds ratio (OR) for any alcohol use disorder was 2.0 for 12-month prevalence and 2.2 for lifetime prevalence. There are interesting and significant differences between comorbidity of anxiety disorders with harmful alcohol use and alcohol dependence. While the comorbidity with alcohol abuse was not increased in this study (ORs 1.0 and 1.1 for 12-month and lifetime prevalence, respectively), the ORs for alcohol dependence were significant (3.1 and 2.8, respectively). Similar results were found also for other drugs of abuse: Data indicate an association of GAD with dependence on nicotine or drugs but not with abuse of nicotine or drugs. The relationship was strongest for drugs (ORs 9.9 and 5.2, respectively). In addition, as demonstrated in previous studies there was also evidence for a strong association of GAD with other mood and anxiety disorders. Further analysis of the database from this study revealed marked sex differences: Men with GAD had significantly higher rates of comorbid alcohol and drug use disorders and reported greater use of alcohol and drugs to help relieve GAD symptoms.

Again, an association of alcohol dependence, but not harmful use, with anxiety disorders (and depression) was shown in a Dutch Study (Boschloo et al. 2011). This study included a sample of 2,329 people with lifetime DSM-IV anxiety disorders or depressive disorders or both and 652 controls. Prevalence rates for alcohol dependence in persons with combined anxiety/depression were 20.3 % (controls: 5.5 %). Prevalence of alcohol abuse was similar in all groups (about 12 %). A number of independent risk factors for alcohol dependence were identified: Male gender, vulnerability factors such as a family history of alcohol dependence or anxiety/depression, childhood trauma, smoking, drug dependence, and early onset of anxiety/depression. There is also some evidence that having a comorbid anxiety disorder is associated with increased substance use severity (Schneier et al. 2010).

The database is far less extensive for drug abuse. In a large national epidemiologic survey in the US, 12-month prevalence estimates were 1.4 % and lifetime prevalence rates 7.7 %, which clearly exceeded the rates of drug dependence (0.6 % and 2.6 %, respectively). Twelve-month prevalence for drug use disorders was associated with any anxiety disorder (OR 2.7); with any PD (OR 3.9; PD with agoraphobia: OR 5.6, without: OR 3.1); with social phobia, OR 2.6; with specific phobia, OR 2.3; and with GAD, OR 4.5. When adjusted for demographic characteristics and other psychiatric disorders, any anxiety disorder was associated with drug use disorders with an OR of 2.1, with drug abuse with an OR of 1.5, and with drug dependence with an OR of 2.8. For drug dependence, GAD had an OR of 2.5. These data correspond in part to studies in alcohol use disorders showing that drug dependence rather than abuse is associated with anxiety disorders, specifically GAD.

The chronological relationship between the onset of anxiety disorders and substance use disorders is complex. Substance-induced anxiety symptoms frequently occur during detoxification from alcohol or benzodiazepines or during intoxication with cannabis or psychostimulants (cocaine, amphetamine) but often improve or vanish within a few weeks. Long-term longitudinal studies in patients with anxiety disorders did not indicate an association of phobias with the onset of alcohol use disorders but rather a modest association between adult subclinical-specific phobias and later-onset alcohol use disorders (OR 3.2); the association was stronger in women than men. Zimmermann et al. (2003) reported 4-year follow-up data from a prospective community survey in 3,021 adolescents. Baseline social phobia and panic attacks significantly predicted subsequent alcohol problems in young adults. This chronological order has recently been confirmed in a longitudinal Dutch study showing that current anxiety disorder significantly predicted first incidence of alcohol dependence (Boschloo et al. 2013).

Flensborg-Madsen et al. (2011) examined the effects of alcohol intake (not alcohol use disorders) on the risk of psychiatric disorders. This prospective cohort study included participants from the Copenhagen City Heart Study (N = 18,146). Participants were followed for up to 26 years. Alcohol intake was measured by self-report, while psychiatric diagnoses were measured through registers. For women, drinking above sensible limits increased the risk for psychiatric disorders in general and especially for anxiety disorders (risk: 2.00). For men, a weekly low to moderate alcohol intake seemed to have a protective effect against developing a psychiatric disorder. Risk for anxiety disorders was lower in men drinking more than 14 drinks per week (OR 0.79). The authors claimed an “apparent protective effect” of alcohol among men as a sign of mental and social well-being and normal functioning.

Grant et al. (2009) studied sociodemographic and psychopathologic predictors of the first incidence of DSM-IV substance use and mood and anxiety disorders by examining data from the WAVE 2 National Epidemiologic Survey on Alcohol and Related Conditions. One-year incidence rates were highest for alcohol abuse (1.02), alcohol dependence (1.70), major depressive disorder (1.51), and GAD (1.12). Incidence rates were greater among men for substance use disorders and greater among women for mood and anxiety disorders, except bipolar disorder, and social phobia. Age was inversely related to all disorders. Interestingly, substance use disorders did not predict any incident mood or anxiety disorder, whereas baseline bipolar I predicted incident drug abuse and baseline PD predicted incident drug dependence. Although these results may be consistent with the self-medication hypothesis, other mechanisms such as shared liability arising from the same genetic or environmental risk factors cannot be excluded (2009).

Melchior et al. (2014) recently reported additional data from the US national Epidemiologic Survey on Alcohol and Related Conditions . Of 34,632 people included, 3.2 % had a diagnosis of lifetime illegal drug use disorder; 21.2 %, a comorbid mood disorder; 11.8 %, a comorbid anxiety disorder; and 45.9 %, comorbid mood and anxiety disorders. In contrast, recent data from the National Comorbidity Survey of a nationally representative sample of the US adult population showed that substance dependence temporally precedes several anxiety disorders, particularly PD (OR 2.62, Goodwin and Stein 2013). The ORs for social phobia (OR 1.7) and agoraphobia (OR 1.78) were smaller. Conversely, the anxiety disorder appeared first in more than 50 % of substance use disorder cases, in nearly 40 % of PTSD cases, and in nearly 30 % of GAD cases. Similarly, a lifetime history of social phobia, PTSD, or GAD significantly predicts lifetime substance dependence (OR 1.51 for social phobia, 2.06 for PTSD, and 1.45 for GAD).

The database for drug use is more limited. Liang et al. (2011) performed a retrospective cohort study on data from the 2007 National Survey of Mental Health and Wellbeing (MHW) in 8,841 adult Australians. Previously, Teesson et al. (2009) had reported data from this study indicating that 19 % of males and 8 % of females with an anxiety disorder had a coexisting substance use disorder and 26 % of males and 11 % of females with an affective disorder had a coexisting substance use disorder. Overall prevalence for drug dependence was 2.56 %. Individuals with an affective disorder or anxiety disorder were at higher risk of harmful use and drug dependence (males: 9.3 %; females: 3.9 %). Again, the self-medication theory or common genetic factors were discussed to explain these findings.

In general, substance use disorders are more prevalent in men, and anxiety disorders are more prevalent in women (Kessler et al. 2005). Women with substance use disorders are more likely to have a comorbid anxiety disorder than men (60.7 % versus 35 %, Kessler et al. 1997) There is some evidence that comorbid anxiety disorders complicate treatment of substance use in women. Farris et al. (2012) studied 260 women treated within an alcohol program and found that lifetime anxiety diagnosis was linked to poorer drinking outcomes post treatment, although women with comorbid anxiety disorders drank less than nonanxious patients before treatment.