One such relevant factor may be anxiety sensitivity (AS), a cognitive vulnerability that has been implicated in a wide range of psychiatric problems, including PTSD (Taylor 1999, 2003). AS is defined as the tendency to fear anxiety-related symptoms due to beliefs that these symptoms will have harmful physical, social, or cognitive consequences (Reiss 1991). Research has consistently demonstrated an association between AS and PTSD (for a review, see Elwood et al. 2009). Further, PTSD appears to be associated with higher levels of AS than all other anxiety disorders with the exception of panic disorder (for which levels of AS are comparable to those observed among individuals with PTSD; Taylor et al. 1992). Stronger evidence for the role of AS in posttraumatic stress comes from studies that utilize a prospective research design. For example, Feldner et al. (2008) found that baseline AS uniquely predicted posttraumatic symptoms at the 12–24 month follow-up in a large nonclinical sample of young adults, above and beyond trait anxiety, baseline posttraumatic stress, and treatment condition (i.e., a brief, computerized AS intervention versus a control condition). Additionally, Fedoroff et al. (2000) found that reductions in AS corresponded with reductions in PTSD symptom severity over the course of cognitive behavioral treatment among survivors of motor vehicle accidents. Further, Marshall et al. (2010) found that AS predicts PTSD symptoms 6 and 12 months post-physical injury. Similar associations between AS and PTSD have also been observed among individuals in SUD treatment. For example, a cross-sectional study of patients in a community-based SUD treatment program found that individuals classified as having PTSD (based on a self-administered questionnaire) reported greater AS than those without PTSD and those with subclinical PTSD symptoms (Bonin et al. 2000). Further, a prospective study of patients receiving treatment for alcohol use disorders found that baseline AS was associated with the severity of PTSD symptoms concurrently and at 1 month follow-up (Simpson et al. 2006).

Another factor relevant to PTSD symptom severity is emotional avoidance (EA), considered to be an underlying mechanism of various forms of psychopathology, including PTSD (e.g., Chawla and Ostafin 2007; Hayes et al. 1996; Salters-Pedneault et al. 2004). EA is subsumed under the broader construct of experiential avoidance (Hayes et al. 1996), which consists of efforts to avoid or alter the form, frequency, or contexts of aversive internal experiences (e.g., thoughts, emotions, and bodily sensations). Although avoidance of trauma-specific thoughts and emotions is already recognized as a core feature of PTSD (American Psychiatric Association 2000), recent theoretical and empirical literature suggests that the generalized avoidance of aversive internal experiences (i.e., experiences that are not necessarily trauma-related) is associated with PTSD as well. For example, Roemer et al. (2001) found that combat veterans with PTSD reported more frequent and extreme withholding of emotions (e.g., concealment of emotions across multiple situations, concealment of all emotional reactions) than non-PTSD controls, and experiential avoidance has been found to be associated with PTSD and depression above and beyond the severity of combat exposure in a sample of veterans presenting for inpatient PTSD treatment (Plumb et al. 2004). Additionally, Marx and Sloan (2002) reported that experiential avoidance mediated the relation between reports of childhood sexual abuse and general psychological distress among female college students. Prospective studies have provided even stronger support for the role of experiential avoidance as a vulnerability factor for the development of post-stressor psychiatric problems. In a series of studies, Plumb et al. (2004) found that experiential avoidance among college students was prospectively associated with greater psychological distress following exposure to stressful life events, and greater PTSD symptom severity following exposure to traumatic events. Likewise, in another prospective study of trauma-exposed college students, Marx and Sloan (2005) found that greater self-reported experiential avoidance at the initial assessment predicted PTSD severity both 4 and 8 weeks later, even after accounting for initial PTSD symptom levels. Despite these interesting findings, however, understanding the role of general emotional avoidance following a traumatic event is complicated by its conceptual overlap with PTSD-specific avoidance. Consequently, avoidance is often a dimension of both the independent and dependent variables in such studies.

In light of these findings, the aim of the current study was to examine AS and EA as potential emotion-related factors associated with PTSD symptom severity among a sample of crack/cocaine dependent patients with a history of traumatic event exposure. Crack/cocaine dependent patients are a particularly relevant population in which to explore these associations, given evidence of high levels of PTSD (Cottler et al. 1992), AS (Lejuez et al. 2006; McDermott et al. 2009), and emotion dysregulation (Fox et al. 2007) within this population. Further, given that previous theoretical and empirical work on the role of AS and EA in other trauma-exposed populations has informed the development of new treatment models (e.g., Hayes et al. 1999; Taylor 1999), understanding the relevance of these constructs to posttraumatic stress within SUD populations may highlight potentially useful targets of intervention among SUD patients with PTSD (an important consideration in light of the negative outcomes associated with this specific diagnostic co-occurrence; Hien et al. 2000; Najavits et al. 1999).

We hypothesized that AS and EA would be associated with PTSD symptom severity above and beyond any shared variance with relevant demographic variables and non-specific anxiety symptoms. In addition, in light of literature suggesting that fear of anxious arousal (i.e., AS) may motivate avoidance behaviors aimed at reducing this emotional distress (i.e., EA; Zvolensky and Forsyth 2002), and that persistent avoidance may, in turn, interfere with the processing of fear-based emotional responses (e.g., habituation, corrective learning; Foa et al. 2006; Foa and Kozak 1986), we further hypothesized that EA would mediate the association between AS and PTSD symptom severity. Specifically, by preventing the emotional processing of fear, EA may be a mechanism through which AS contributes to greater PTSD symptom severity (Taylor 2003). According to this perspective, it is pervasive efforts to avoid aversive emotions, rather than the emotions themselves, that leads to the development and maintenance of psychopathology, such as PTSD (Hayes et al. 1996). We are not aware of any studies examining this meditational model as a predictor of posttraumatic stress; however, previous research has found that experiential avoidance mediates the association between dimensions of AS and various other psychiatric and behavioral problems, including depression (Tull and Gratz 2008), borderline personality disorder (Gratz et al. 2008), and coping-motivated drinking (Stewart et al. 2002). Although the cross-sectional nature of the current study precludes a true test of mediation (which requires a specific temporal relationship among the variables in question, such that the mediator must occur after predictor and before the outcome; Kraemer et al. 2002), it does provides a preliminary test of this meditational model, similar to that provided in the studies cited above (Gratz et al. 2008; Stewart et al. 2002; Tull and Gratz 2008).

Regardless of the directionality of these relationships, the tendency to engage in EA may have important implications for the treatment of individuals with SUD-PTSD. For example, although reducing trauma-related avoidance is a primary target of cognitive behavioral treatments for PTSD (Foa et al. 2007; Resick and Schnicke 1996), our results suggest that avoidance may extend to a fuller range of emotional experiences, consistent with the findings of Roemer et al. (2001). This tendency may have a negative impact on patient health, as studies have found that the purposeful suppression of emotional responses is associated with paradoxical increases in acute physiological arousal and emotional distress (e.g., Campbell-Sills et al. 2006; Gross and Levenson 1997), as well as lower long-term functioning and well-being (Gross and John 2003). Additionally, both AS and EA have previously demonstrated associations with negative SUD treatment outcomes. For example, AS was found to be prospectively associated with SUD residential treatment drop-out among heroin and crack/cocaine users (Lejuez et al. 2008). Similarly, avoidance of emotional distress during behavioral persistence tasks was associated retrospectively with self-reported duration of abstinence (Daughters et al. 2005) and prospectively with early treatment drop-out (Daughters et al. 2005). Thus, interventions focused on decreasing AS and EA among trauma-exposed SUD patients may help mitigate the negative health- and treatment-related outcomes found among SUD patients with PTSD.

Although interesting, our findings must be considered in light of the limitations present. As previously noted, these correlational and cross-sectional data preclude identification of the precise nature or direction of the associations between these variables. In addition to the potential reciprocal relationship described above, it may be that AS and EA have an interactive effect on PTSD.¹ Longitudinal data are necessary to begin addressing these issues (e.g., Marshall et al. 2010). Second, these results must be interpreted with caution due to limited statistical power resulting from a small sample size. Third, whereas AS is an anxiety-specific vulnerability factor, EA is not specific to fearful or anxious responding. Thus, although EA theoretically captures avoidance of a variety of emotional experiences (including anxiety), a measure of anxiety-specific avoidance may prove to be a more powerful mediator of the association between AS and PTSD. Future research should test this possibility, as well as whether EA mediates the relation between other aversive emotions (e.g., guilt) and PTSD symptom severity. Further, given that data were collected from a primarily African American sample of crack/cocaine dependent patients in residential SUD treatment, findings may not be generalizable to individuals with co-occurring SUD and PTSD symptoms in general, or to those with specific SUDs (e.g., alcohol or heroin). For example, it was surprising that substance use severity was not correlated with PTSD symptom severity, a finding that may reflect factors specific to the current sample. Alternatively, this finding may be due to the differential reference periods of the assessment instruments. Whereas substance use severity was assessed retrospectively over the past year, PTSD symptoms were assessed retrospectively over the past month. It is possible that similar reference periods would have resulted in a stronger correlation between measures. Thus, it will be important for future studies to examine relations between AS, EA, and PTSD symptom severity in other SUD populations, and to assess both past and present self-reported PTSD symptoms.

In addition, one issue that may have influenced our assessment of EA is the degree to which participants were capable of attending to, identifying, and labeling emotions (i.e., emotional awareness and clarity). Recent theoretical and empirical work highlights the important role these processes play in effective emotion regulation (e.g., Gratz and Roemer 2004), and suggests that lack of emotional clarity may be greater among individuals reporting high rates of PTSD symptoms (Tull et al. 2007). Given that the reliable and valid assessment of EA will likely depend, in part, on an individual’s awareness and clarity of emotional experiences (Tull et al. 2008), future research should attempt to account for this variability. The inclusion of behavioral measures of EA and related constructs (see Gratz et al. 2006) may have utility in addressing this limitation.

Future research is also needed to examine how AS and EA may contribute to a functional relationship between PTSD and SUD. In line with a self-medication model (Brady et al. 2004), previous research suggests that a worsening of trauma-related symptoms in individuals with SUD-PTSD may lead to an increase in substance cravings and usage (Back et al. 2006; Saladin et al. 2003). Further, there is growing evidence that substance use may function as a way for individuals with PTSD to modulate or avoid aversive emotions in general, not just those related to traumatic event exposure (Ouimette et al. 2007; Waldrop et al. 2007). Greater understanding of this functional relationship would likely be aided by further identification of the cognitive, emotional, and behavioral vulnerabilities associated with PTSD among individuals with SUD. For example, preliminary evidence suggests that several dimensions of emotion dysregulation (McDermott et al. 2009), including emotional nonacceptance (Gratz et al. 2007), may be useful targets for future research.