Anxiety Disorders, Comorbid Substance Abuse, and Benzodiazepine Discontinuation: Implications for Treatment

David H. Barlow

INTRODUCTION

Comorbidity among various disorders complicates research and practice. Comorbidity among emotional disorders and substance use disorders is a particularly thorny problem due to the dearth of relevant clinical research.

This chapter reviews what is known about the comorbidity of substance use and anxiety disorders and presents recent data collected in the context of comorbid anxiety and mood disorders that may have implications for the relationship of anxiety and substance use disorders.

The specific case of the relationship of benzodiazepine use to successful outcome of psycho- social treatments and recent developments in successful psychosocial strategies for discontinuing benzodiazepines in anxious patients may provide important information for future studies. This chapter begins with a brief review of data on the co-occurrence of substance use and anxiety disorders.

COMORBIDITY AMONG SUBSTANCE USE AND ANXIETY DISORDERS

A number of studies have reported a high rate of comorbidity among anxiety and substance use disorders. Most of these studies have surveyed alcohol dependence and abuse.

Rates of comorbidity have typically been calculated in two different ways. First, the prevalence of anxiety disorders has been examined in alcohol dependence and abuse patient samples.

Second, rates of alcohol dependence and abuse have also been examined in samples of outpatients with anxiety disorders. The majority of surveys have followed the first approach and have found that the lifetime prevalence of clinically significant anxiety disorders in patients with alcohol abuse and dependence ranges from 25 percent to 45 percent for patients with clearly defined anxiety disorders, but may approach 60 percent if one includes identifiable anxiety disorders that are subthreshold in terms of severity (Bowen et  al. 1984; Chambless et al. 1987; Hesselbrock et al. 1985; Mullaney and Trippett 1979; Smail et al. 1984; Cox et al. 1989; Johannessen et al. 1989).

Surveys using the second approach and examining rates of alcohol dependence and abuse in anxiety disorder outpatient samples suggest that approximately 15 percent to 25 percent present with evidence of current or past alcohol abuse or dependence (Bibb and Chambless 1986; Thyer et al. 1986). Himle and Hill (1991) found that the frequency of alcohol abuse or dependence differed among persons with various anxiety disorders.

For example, the percentage of alcohol abuse or dependence among those individuals with a principal diagnosis of panic disorder (PD) with agoraphobia (who may also have presented with additional anxiety disorders) was 31.5 percent, as compared to 24.6 percent for obsessive-compulsive disorder and 14.4 percent for a specific phobia.

Thus, it would seem that some anxiety disorders confer a higher risk for substance abuse then others. In any case, there is evidence that patients presenting with these comorbid pictures have more clinically severe conditions than individuals with either condition alone.

Thus, there are reasons to examine factors contributing to comorbidity in this subgroup more closely. One method of examining the possible reasons for the acquisition of comorbid disorders is to ascertain a temporal sequence in their onset. Most studies indicate that anxiety precedes alcohol abuse and dependence.

This pattern would seem to confirm the frequent clinical observation that many individuals with anxiety disorders begin to abuse alcohol with the purpose of self-medicating their anxiety disorders. However, Kushner and colleagues (1990) noted that the pattern seems to hold true only for some disorders, such as PD with or without agora-phobia, social phobia, and specific phobia.

For some other disorders, particularly generalized anxiety disorder (GAD) and depression, the more prevalent pattern may be the reverse; that is, substance abuse seems to contribute to the onset of GAD and depression. One possible mechanism of action here is that the individual experiences a loss of control over the substance use subsequent to addiction and develops reactive anxiety or depression.

Illicit drug use has also been reported to precipitate anxiety disorders. For example, Aronson and Craig (1986), as well as Louie and colleagues (1989), reported a number of cases in which cocaine use and/or withdrawal from cocaine precipitated panic attacks.

In these  cases the resulting panic disorder continued well after the cessation of cocaine use. In fact, as many as 30 percent of patients presenting with PD have reported an onset associated with either licit or illicit drug use (Barlow 1988), with marijuana being one of the more common precipitants.

Hyperventilation and other symptoms associated with withdrawal from alcohol have also been reported to trigger longlasting PD (Weissman 1988). In cases where substance abuse seems to "trigger" anxiety disorders, clinical strategies might target the substance use first before addressing related anxiety on the chance that anxiety, to the extent that it might be related to the substance use, would concurrently remit.

These clinical speculations, however, are nothing more than assumptions since little is known about the effects of targeting one disorder when treating additional comorbid disorders in an individual.

COMORBIDITY AMONG ANXIETY AND MOOD DISORDERS: IMPLICATIONS FOR COMORBID SUBSTANCE USE DISORDERS
Research from the author’s anxiety disorders research clinic has produced some evidence on the effects of comorbidity among anxiety and mood disorders on treatment outcome, both short and long term. Since these results are somewhat surprising, it is possible that they may have some implications for similar comorbid patterns among anxiety disorders and substance use disorders.

One recently analyzed set of data examined the impact of treatment for panic disorder using an effective cognitive-behavioral treatment (Barlow et al. 1989) on the course and outcome of generalized anxiety disorder that was not directly treated (Brown and Barlow 1992). GAD was chosen because it is the most frequently co-occurring diagnosis in patients with a principal diagnosis of PD (Moras et al., submitted).

For purposes of this analysis, the comorbid presence of GAD was considered at both a clinical level of severity as well as a subclinical level of severity in which GAD was clearly identifiable but was not considered severe enough to interfere substantially with functioning. As noted in figure 1, of 68 panic disorder patients treated, 32 percent had a clinically significant GAD additional diagnosis at pretreatment, with an additional 9 percent evidencing subthreshold GAD.

At posttreatment the rate of GAD above threshold declined to 9 percent, whereas subthreshold GAD increased to 16 percent because several patients with a clinically significant GAD at pretreatment moved to the subclinical category at posttreatment. These results were relatively stable at a 3-month followup.

Thus, in this example, a comorbid disorder improved with successful treatment of the target disorder in spite of the fact that no attempts were made to treat it directly. Of course, one possible reason for these results is that GAD and PD share many symptoms, with GAD often considered to be the "basic" anxiety disorder (Brown et al. 1994).

Thus, the success-ful treatment of panic disorder may have "generalized" to symptoms comprising GAD such as anxious arousal and cognitions of future danger. 

 ~~~
Continued on source document:
NIDA Research Monograph, Number 172