Anxiety and substance use

Grant and colleagues' analyses of the National Epidemiologic Survey on Alcohol and Related Conditions consistently find that anxiety disorders co-occur with alcohol use disorder at roughly two to three times the rate expected from base rates alone. The overlap is present in every specific anxiety disorder studied, with somewhat higher odds ratios for generalized anxiety disorder and social anxiety disorder. NSDUH tracks the same pattern for a range of substances.

It runs both ways. The self-medication hypothesis, articulated by Khantzian and refined in decades of subsequent work, holds that some people use alcohol or other substances to reduce anxiety in the short term. Longitudinal cohorts, including work by Kushner and colleagues on alcohol and anxiety, support this pattern for some patients. At the same time, alcohol and benzodiazepine withdrawal states produce anxiety and panic-like symptoms directly, so heavy or repeated substance use can drive new-onset anxiety symptoms. In an individual case the history usually points to one direction or the other, but survey data alone can't distinguish them at the population level.

Cannabis, stimulants (including prescribed and illicit), and sedatives all co-occur with anxiety at elevated rates. The cannabis picture is mixed: some users report anxiolytic effects, some develop cannabis-induced anxiety, and heavy chronic use is associated with panic. Stimulants can produce or worsen anxiety through their direct pharmacology and through the stress of withdrawal. Sedatives, especially benzodiazepines, produce anxiety on withdrawal that is often misattributed to the underlying disorder returning.

The older model treated substance use first and anxiety later, or the reverse. Randomized comparisons of integrated versus sequential care generally favor integrated approaches, and both SAMHSA (Treatment Improvement Protocol 42) and the American Psychiatric Association support treating co-occurring conditions together. Practically, this can mean combining medication for anxiety with a substance-use-focused therapy, or delivering CBT skills that address both. Benzodiazepine prescribing for anxiety in patients with alcohol or opioid use disorder is generally avoided because of overdose and dependence risks (see the evidence summary on benzodiazepines for the safety context).

Anxiety disorders and substance use disorders co-occur at high rates. National survey data, including the SAMHSA National Survey on Drug Use and Health and the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), consistently show that people with an anxiety disorder are more likely than the general population to also meet criteria for a substance use disorder, and the reverse is also true. Grant and colleagues' analyses of NESARC-III put the odds of comorbid alcohol use disorder in adults with a past-year anxiety disorder at roughly two to three times that of adults without anxiety, depending on which anxiety disorder is being examined.

The relationship runs in more than one direction. For some people, alcohol or other substances are used to manage anxiety symptoms in the short term, a pattern sometimes called self-medication, which can develop into a substance use disorder over time. For others, heavy substance use, and especially withdrawal, can produce or worsen anxiety. Alcohol withdrawal in particular can produce intense anxiety and panic-like symptoms. Untangling which came first is often not possible from survey data.

Comorbid anxiety and substance use is associated with a harder course than either condition alone. People with both conditions have higher rates of relapse, more functional impairment, and higher healthcare utilization. Clinical guidance increasingly favors treating both conditions together rather than insisting one be resolved before the other is addressed.

Survey data shows the association clearly but can't establish the direction of cause. The two conditions influence each other, and a single person's history may run either way.

The reported overlap depends on how the survey defines substance use disorder and which time window is used.