Assessing shared psychological constructs as risk factors in comorbid PTSD-AUD combat-exposed male veterans

ABSTRACT

PTSD and AUD are frequently comorbid post-trauma outcomes. Much remains unknown about shared risk factors as PTSD and AUD work tends to be conducted in isolation. We examined how self-report measures of distress tolerance (DT), experiential avoidance (EA), and drinking motives (DM) differed across diagnostic groups in white, male combat-exposed veterans (n = 77). A MANOVA indicated a significant difference in constructs by group, F (5, 210) = 4.7, p = <.001. Follow-up ANOVAs indicated DM subscales (Coping: F (3,82) = 21.3; Social: F (3,82) = 13.1; Enhancement: F (3,82) = 10.4; ps = <.001) and EA (F (3,73) = 7.8, p < .001) differed by groups but not DT. Post hoc comparisons indicated that mean scores of the comorbid and AUD-only groups were significantly higher than controls for all DM subscales (all ps < .01). EA scores were significantly higher for the comorbid as compared to control (p < .001) and PTS-only (p = .007) groups. Findings support shared psychological factors in a comorbid PTSD-AUD population.

What is the public significance of this article?— This study indicates constructs that may be important in the common co-occurrence of AUD and PTSD in trauma-exposed veterans. The co-occurrence of PTSD-AUD leads to higher symptom severity and difficulty in treating these disorders and efforts to identify relevant factors as identified in this study can point to future research and targets for intervention and prevention.

Introduction

Trauma exposure is a transdiagnostic risk factor that is associated with increased risk for alcohol use disorder (AUD), posttraumatic stress disorder (PTSD), and comorbid PTSD-AUD (Cerdá et al., 2011; Guinle & Sinha, 2020). The frequent comorbidity of PTSD-AUD is well-established (e.g., Debell et al., 2014; Kachadourian et al., 2014; Pietrzak et al., 2011) and particularly frequent in veteran populations (e.g., Blakey et al., 2022; Jakupcak et al., 2010). In a nationally representative sample of US veterans with probable AUD, 20.3% also had probable PTSD; among those with probable PTSD, 16.8% had probable AUD (Norman et al., 2018). The comorbidity of PTSD-AUD is clinically impactful (i.e., greater symptom severity, higher rates of suicidal ideation, worse cognitive and physical functioning, faster alcohol relapse, higher treatment dropout) compared to either disorder alone (e.g., McCauley et al., 2012; Ouimette et al., 2006; Read et al., 2004). The high rates of comorbidity suggest that AUD and PTSD likely possess shared underlying risk factors. However, much remains unknown regarding shared etiological underpinnings and risk factors involved in their comorbidity.

Although models of phenotypic causality such as the self-medication model (i.e., individuals use alcohol to dampen aversive symptoms of PTSD) and the susceptibility model (i.e., alcohol use predisposes to higher risk for trauma and subsequently PTSD) are the most frequently studied (Hawn et al., 2020), another plausible model is the common factors model. The common factors model suggests that shared factors may contribute to an increased risk for, as well as maintenance or exacerbation of, both conditions (Danovitch, 2016). These shared risk factors include common genetic factors (e.g., Sartor et al., 2011; Sheerin et al., 2020; Wolf et al., 2010; Xian et al., 2000), environmental experiences (e.g., trauma exposure), and personality or psychological traits (e.g., impulsivity, distress tolerance, coping methods; Lanius et al., 2011) that may contribute to risk for both disorders. However, much remains unknown about the impact of these shared risk factors in PTSD-AUD comorbidity, as most research is conducted on these conditions separately (i.e., one condition is explicitly excluded or overlooked in research on the other), resulting in a dearth of work on comorbid PTSD-AUD samples. It remains an empirical question as to whether AUD-PTSD comorbidity represents simply greater severity compared to either disorder alone or whether there are nuanced relationships among risk factors for each disorder and their comorbid presentation. Therefore, examination of potential unique patterns of association may improve our understanding of shared risk factors. There are numerous potentially shared psychological risk factors of PTSD-AUD across numerous domains; here, we focused on the negative valence and emotionality domain (Boness et al., 2021) and assessed how PTSD-AUD compares to either disorder alone through measures of distress tolerance, drinking motives, and experiential avoidance. Although these factors likely share some aspects, given presence within the same domain, they also capture unique features relevant for PTSD and/or AUD.

Distress tolerance

Distress tolerance (DT) is the perceived capacity to experience and withstand negative physical and psychological states (Leyro et al., 2010). DT has clinical implications as a risk or protective factor for substance use and mental health disorders (Holliday et al., 2016; Vujanovic et al., 2017). The examination of DT with alcohol use outcomes has found that lower levels of DT are related to greater alcohol use and craving (e.g., Khan et al., 2018; Paltell & Berenz, 2022; Zegel et al., 2019). Yet, more work on DT related to AUD, in particular, is needed. A more extensive literature examines DT and PTSD, demonstrating consistent, negative associations (i.e., DT is negatively associated with PTSD symptom severity; Akbari et al., 2022; Zegel et al., 2019). Despite these promising possible associations for each condition alone, few studies evaluate DT as a potential underlying risk factor for the comorbidity of PTSD-AUD. Of the limited literature available, the results are mixed, with some demonstrating a moderating role of DT in the association of alcohol use severity with PTSD symptom severity (Vujanovic et al., 2011; Zegel et al., 2019), while other studies found no association between DT and alcohol outcomes when PTSD was included (Ranney et al., 2021). Given these conflicting findings, continued research is needed (Paltell & Berenz, 2022).

Drinking motives

Drinking motives (DM) represent motivations to drink to attain subjectively valued outcomes such as enhancement (i.e., drinking to increase positive emotions; Cooper et al., 1992), social (i.e., drinking to facilitate social interactions; Sjödin et al., 2021), and coping (i.e., drinking to cope with negative affect; Grant et al., 2007). A well-established literature identifies greater levels of all drinking motives as risk factors for greater alcohol consumption, higher levels of alcohol dependence, and adverse consequences of drinking (Bresin & Mekawi, 2021; Cheng et al., 2017). However, when evaluating drinking motives in comorbid PTSD-AUD, the literature to date focuses more on the self-medication model, which describes using drinking to cope with emotions, stressors, and mental health symptoms such as PTSD symptoms. Research in veteran samples has found that those with PTSD, compared to those without, were more likely to endorse coping motives for increased alcohol use (McDevitt-Murphy et al., 2015). Miller et al. (2017) expanded on this work, demonstrating coping with drinking was the strongest mediator between increased PTSD symptom severity and increased alcohol use in young male veterans. A more limited literature on other motives has found significant associations between PTSD symptom severity with enhancement and social motives for alcohol use. PTSD symptoms have been found to be positively associated with greater average alcohol quantity using enhancement motives in males and females (Lehavot et al., 2014). In heavy-drinking veterans with PTSD, social motives were significantly correlated with adverse alcohol consequences (McDevitt-Murphy et al., 2015). Further investigation into DM as a shared risk factor for PTSD-AUD comorbidity is needed, particularly expanding into other motives (i.e., enhancement, social) to examine its role in alternative models of comorbidity, such as the common factors model.

Experiential avoidance

Experiential avoidance (EA) is a maladaptive emotion regulation strategy used to reduce the frequency or intensity of internal experiences via avoidance, suppression, or other forms of control (Levin et al., 2012; Warnke et al., 2018). EA has been identified as a risk factor for amplifying distress in the context of trauma (Cobb et al., 2017). Previous research has demonstrated that following trauma exposure, engagement in EA is related to PTSD symptom severity over time (e.g., Hetzel-Riggin & Meads, 2016; Marx & Sloan, 2002, 2005). To date, there is limited literature on EA in samples with AUD or comorbid PTSD-AUD. However, one study found EA to be significantly associated with alcohol-related problems, and another found EA to be a possible facilitator for the association between posttraumatic stress symptoms and AUD among combat veterans (Feingold & Zerach, 2021; Levin et al., 2012). Although this limited work suggests the relevance of EA for PTSD, AUD, and their comorbidity, further research is needed to confirm these initial findings.

The present study examined these psychological constructs (i.e., distress tolerance, drinking motives, experiential avoidance) as potential shared risk factors underlying comorbid PTSD-AUD. Further, given the conceptual overlap, at least in part, of these constructs, we examined these constructs together. We examined group differences in these constructs in a sample of combat trauma-exposed veterans within four groups: current posttraumatic stress symptoms only (PTS-only); current AUD only (AUD-only); current PTS and AUD (comorbid); and combat trauma-exposed controls (control). Based on the existing literature in this area, we hypothesized that the comorbid group would endorse greater DM and EA and lower DT compared to all groups. Second, we hypothesized that PTS-only and AUD-only groups would endorse greater DM and EA and less DT compared to the trauma-exposed controls.

Methods

Participants came from a larger parent study examining shared genetic vulnerability and fear-learning processes underlying PTSD-AUD comorbidity. Participants were recruited through the local Veterans Affairs Medical Center (VAMC) by advertising (e.g., flyers, direct mail). Following phone screens for initial eligibility, and then informed consent for those who qualified, participants engaged in a series of structured clinical diagnostic interviews and then completed a battery of self-report questionnaires via REDCap (Harris et al., 2019). All study activities were conducted in accordance with the Declaration of Helsinki and approved by the local university and VAMC Institutional Review Boards. Given the diagnostic group design for the laboratory task and examination of genetic risk, inclusion criteria for the parent study included European ancestry; male; ability to provide informed consent; confirmation of deployment and combat exposure; some current alcohol use; and age between 21 and 45 years. Exclusion criteria for the parent study assessed during the initial eligibility phone screening included the presence of psychosis or mania; unstable prescription medication or current benzodiazepine use; current major depressive disorder (MDD) per self-report and being above the threshold of 10 on the Patient Health Questionnaire-9 (Kroenke et al., 2001); and current substance use disorder (SUD; excluding alcohol, nicotine, and caffeine) per self-report and being above a threshold of 6 on the Drug Abuse Screening Test (DAST; Villalobos-Gallegos et al., 2015); and history of a moderate or severe traumatic brain injury as assessed by the Traumatic Brain Injury Screener (Schwab et al., 2007). Of the 576 eligibility phone screens conducted, 20.1% qualified, 67.5% disqualified, and 12.3% were lost to follow-up prior to full eligibility determination. Of the N = 114 who qualified, N = 103 were consented and enrolled. The present sample (N = 77) of white, male combat trauma-exposed veterans represents those who had confirmed Criterion A combat trauma exposure and completed the clinical interview and self-report measures, categorized into specific diagnostic groups.

Measures

Clinical interviews

The Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998)

The MINI is a short, structured clinical interview for psychiatric disorders listed in the DSM-5. Questions are designed to be answered in a yes/no fashion, and current (past year) disorder status was focused on in the present study. The MINI has excellent inter-rater reliability and test–retest reliability (all kappa values above 0.75); good sensitivity, with a value of 0.70 or greater for all but three disorders (dysthymia, OCD, and drug dependence); and excellent specificity (above 0.85; Sheehan et al., 1998). In the present study, the MINI was used to determine AUD status as well as the presence of other comorbid conditions.

Clinician-Administered PTSD Scale (CAPS-5; Weathers et al., 2018)

The CAPS-5 is a 30-item structured diagnostic interview used to measure a current diagnosis of PTSD, a lifetime diagnosis of PTSD, and assess the severity of PTSD symptoms. The CAPS-5 has excellent internal consistency (Cronbach’s alpha = .88; Weathers et al., 2018). The CAPS-5 past month was assessed for combat trauma. History of exposure to combat or a warzone was identified before the interview using the Life Events Checklist (LEC) and confirmed to meet Criterion A as part of the CAPS-5.

Diagnostic group determination

Participants were categorized into four study groups based on the clinical interviews. Any AUD (i.e., mild or more) was considered AUD positive. Individuals meeting current PTSD diagnostic status, as well as subthreshold PTSD, were considered PTS positive. Subthreshold PTSD (N = 4) was defined as having at least three of four symptom clusters (B-E) met. Participants in the PTS-only group were PTS positive on the CAPS-5 and AUD negative on the MINI, while participants in the AUD-only group were AUD positive and PTSD negative. The comorbid group participants were positive for PTS and AUD, while trauma-exposed controls were negative for both. Trauma-exposed controls were determined based on endorsement of the LEC as noted above and meeting Criterion A per the CAPS but not meeting PTSD or sub-threshold criteria. Subthreshold PTSD participants (N = 4) did not significantly differ from PTSD participants on any of the study measures and were thus included in the PTS group to increase the sample size. Current control participants who had lifetime PTSD or subthreshold PTSD (N = 6) did not significantly differ from the other control participants on any of the study measures and were included in the control group to increase the sample size. See Supplemental Table 1 for comparisons.

Self-report measures

Distress tolerance scale (DTS; Simons & Gaher, 2005)

The DTS is a 15-item self-report scale used to measure individual-level DT. Participants are asked to rate their beliefs surrounding negative affect on a 5-point Likert scale whereby “1” = strongly agree and “5” = strongly disagree. Single, higher-order DT factor scores have good internal consistency (Simons & Gaher, 2005). The sum of all items, with the higher the sum of all items, the higher that individual’s distress tolerance was used. In the current study, internal consistency was good (Cronbach’s alpha = .91).

Drinking motives questionnaire (DMQ; Cooper etal., 1992)

The DMQ is a 20-item self-report scale used to measure the frequency of three types of drinking motives: enhancement, social, and coping. Each item is rated on a 4-point Likert Scale with a score of “1” indicating “never/almost never” and a score of “4” indicating “always/almost always.” Scoring used in the current study was the sum of each subscale, with higher scores in a subscale indicating higher levels of motivation to drink for that subscale. The DMQ has been found to have good internal consistency among the three subscales (Cronbach’s alpha = .77–.85; Cooper et al., 1992). The current study’s internal consistency was good (Cronbach’s alpha = .83–.90).

Multidimensional experiential avoidance questionnaire (MEAQ; Gámez et al., 2011)

The MEAQ is a 62-item self-report measure that assesses experiential avoidance using total score and by subscales. Each item is rated on a 6-point Likert Scale with a score of “1” indicating “strongly disagree” and a score of “6” indicating “strongly agree.” The scoring used in the current study was the sum of all items, with high scores indicating more experiential avoidance. There are six subscales within the MEAQ: behavioral avoidance, distress aversion, procrastination, distraction and suppression, repression and denial, and distress endurance. These were examined in follow-up analyses. The MEAQ has been found to have good internal consistency and good convergent and criterion validity, with Cronbach’s alpha = .86 overall (Gámez et al., 2011). In the current study, internal consistency for the overall sum was good (Cronbach’s alpha = .92), as well as the internal consistency of subscales (Cronbach’s alpha = .81 to .92), except for the distress endurance subscale (Cronbach’s alpha = .70).

Data analytic plan

Before conducting primary analyses, all data were inspected for normality, linearity, outliers, homogeneity of variance-covariance matrices, and multicollinearity; no serious violations were noted. A between-subjects multivariate analysis of variance (MANOVA) was conducted to investigate differences across diagnostic-based study groups by the five measures of interest (DT, DM-Cope, DM-Social, DM-Enhancement, EA). A MANOVA was selected over separate univariate analyses of variance (ANOVA) to start, given the interrelationship between these constructs overall and in this sample (see correlations in Table 1 demonstrating a range of association, some strong, but not unity). Significant findings were then followed up with ANOVAs and post hoc Tukey’s HSD test comparisons, or Games-Howell where variances were unequal, to identify significant differences between the specific diagnostic-based study groups and the measures of interest. Finally, in order to examine whether other potentially relevant factors, specifically depressed mood and combat severity, may have contributed to differences among the groups, follow-up ANCOVAs were conducted using the Beck Depression Inventory-II (BDI-II; Wang & Gorenstein, 2013) and the combat experiences scale of the Deployment Risk and Resilience Inventory-2 (DRRI-2; Vogt et al., 2013). Analyses were conducted using SPSS (version 29; IBM Corp, 2023). We note that sensitivity analyses were conducted removing the subthreshold PTSD participants (n = 4) and the control individuals with lifetime PTSD/subthreshold PTSD (n = 6) from the MANOVA. All patterns remained the same. Therefore, results present the full, combined sample groups.

Table 1.

Correlations and Means/SDs among psychological constructs in full sample and clinical groups.

						Total Sample (n = 77)	Comorbid (n = 14)	PTSD only (n = 15)	AUD only (n = 16)	Control (n = 32)
Variable	1	2	3	4	5	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
1. DTS	–					3.69 (0.78)	3.21 (0.75)	3.59 (0.64)	3.69 (1.08)	4.00 (0.63)
2. DMQ-Coping	−.44**	–				8.09 (3.20)	11.81 (3.39)	7.31 (2.30)	9.16 (3.10)	6.17 (1.38)
3. DMQ-Social	−.08	.52**	–			9.97 (3.23)	11.69 (3.03)	8.63 (2.42)	12.53 (3.26)	8.40 (2.32)
4. DMQ-Enhance	−.20	.80**	.64**	–		8.52 (3.49)	11.19 (3.87)	7.88(3.44)	10.16 (3.78)	6.71(1.64)
5. MEAQ	−.72**	.35**	.21	.15	–	182 (42.67)	213.29 (28.22)	199.93 (47.36)	179.25 (35.79)	160.41 (37.61)

*p < .05; **p < .01. DTS = Distress Tolerance Scale; DMQ = Drinking Motives Questionnaire; MEAQ = Multidimensional Experiential Avoidance Questionnaire.

Results

Our sample consisted of 77 combat trauma-exposed white male veterans. The sample had a median age of 36, a median of 2 deployments, and primarily served in the Army (see Table 2 for complete sample characteristics). As determined by the CAPS-5 and MINI, our diagnostic groups consisted of 16 in AUD-only, 15 in PTS-only, 14 in comorbid, and 32 in control. Sample means and correlations among study constructs are presented in Table 1.

Table 2.

Demographic characteristics of study sample (N = 77).

Characteristic	N (%)
Number of Deployments
0	1 (1.3)
1	26 (33.3)
2	30 (38.5)
3	9 (11.5)
4	7 (9.0)
5 or more	4 (5.1)
Military branch
Army	49 (62.8)
Navy	4 (5.1)
Marine Corps	15 (19.2)
Air Force	6 (7.7)
Coast Guard	0 (0)
Reserves	3 (3.8)
Education
Less than high school	0 (0)
High school graduate/GED	4 (5.1)
Some college	21 (26.9)
College graduate	37 (47.4)
More than college	15 (19.2)
Income
Less than $20,000	8 (10.3)
$20,000–$50,000	16 (20.5)
$50,000–$80,000	24 (30.8)
Over $80,000	24 (20.8)
Decline to answer	5 (6.4)
Marital status
Never married	21 (26.9)
Separated/divorced	7 (9.0)
Currently married or cohabitating	49 (62.8)

A MANOVA conducted with the four clinical groups predicting mean scores for the psychological constructs (Table 3) revealed a statistically significant large-size multivariate effect for the psychological constructs (with the exception of a significant but small effect for DT) by diagnostic groups. When examining the dependent variables separately, the clinical groups had a significant main effect on all psychological constructs. Follow-up ANOVA models were conducted on the significant associations.

Table 3.

MANOVA results of psychological constructs.

	df	F	p	ηp2
Overall Model	15	5.15	<.001	0.269
Construct
DTS	3	3.15	.03	0.101
DMQ-Cope	3	18.46	<.001	0.542
DMQ-Social	3	16.25	<.001	0.513
DMQ-Enhancement	3	10.52	<.001	0.401
MEAQ	3	7.57	<.001	0.209

DTS = Distress Tolerance Scale; DMQ = Drinking Motives Questionnaire; MEAQ = Multidimensional Experiential Avoidance Questionnaire.

Means and SDs by clinical groups from the ANOVA models are presented in Table 1. The ANOVA for the DM subscales revealed a statistically significant difference between all DM subscales scores by the diagnostic groups; Coping: F (3,82) = 21.3, p = <.001, partial eta squared = .44; Social: F (3,82) = 13.1, p = <.001, partial eta squared = .32; Enhancement: F (3,82) = 10.4, p = <.001, partial eta squared = .28. Post-hoc comparisons of the subscales indicated that the mean scores in the comorbid group and the AUD-only group were significantly higher than the control group for all three DM subscales (p < .01 for all comparisons). The comorbid group and the AUD-only group mean scores for the social subscale were also significantly higher than the PTS-only group (p < .05; p < .001, respectively). The comorbid group mean score for the coping subscale was also significantly higher than the PTS-only group (p < .01). All other comparisons were nonsignificant.

Follow-up ANOVA analyses for EA demonstrated a statistically significant difference by the diagnostic groups: F (3,73) = 7.8, p < .001, partial eta squared = .24. Post-hoc comparisons indicated that the comorbid and PTS group’s mean scores on the MEAQ were significantly higher than the control group (comorbid: p < .001; PTS: p = .007); no other comparisons were significant. Based on the significant results of the total MEAQ score, follow-up exploratory analyses were conducted (see Supplemental Table 2) to examine the subscales, as done in prior work (Gámez et al., 2011).

In the follow-up ANOVA analyses for DT, there was a statistically significant difference between the total DT score by the diagnostic groups: F (3,75) = 4.1, p < .01, partial eta squared = .14. Post-hoc comparisons indicated that the comorbid group’s mean score was significantly lower (M = 3.23, SD = 0.74) than the control group (M = 3.98, SD = 0.64). However, no other significant group differences were found.

It is noted that the presence of potentially meaningful, yet smaller effects (i.e., partial eta squared values of .10 or less) between groups may not have been detected given cell sizes and resulting impacts on power. In follow-up ANCOVA analyses, combat exposure severity and depression symptoms were examined. Combat exposure was not significantly associated with groups (p = .06) and did not change overall findings. However, after adjusting for depression symptoms, there were no longer significant group differences on DTS or MEAQ (p = .98, p = .44, respectively). There was a small, significant association between depression with DTS and with MEAQ (partial eta squared = .18, partial eta squared = .23, ps = <.001). Group differences found in the original MANOVA remained significant for DMQ-Cope (partial eta squared = .35), DMQ-Social (partial eta squared = .39), and DMQ-Enhancement (partial eta squared = .27).

Discussion

The present study examined a set of purported underlying shared psychological risk factors of DT, DM, and EA for comorbid PTSD-AUD in a non-treatment-seeking combat-exposed male veteran sample. We hypothesized that the comorbid group would demonstrate the highest endorsement of DM and EA and lowest endorsement of DT compared to all groups, and the PTS-only and AUD-only groups would endorse DM, EA, and DT at levels between those of the comorbid group and the trauma-exposed controls. Findings from the overall MANOVA demonstrated that when considering all constructs’ effects, the clinical groups differed by DM subscales and EA, but not DT. It is noted that results differed when examined combined versus separately, highlighting the importance of examining correlated constructs together to identify which are likely more relevant. Post-hoc analyses of DM subscales and EA partially confirmed our hypotheses. The comorbid group, as hypothesized, generally demonstrated greater severity in all DM subscales and in EA, but only compared to some groups. Our hypotheses regarding the PTS-only and AUD-only groups were generally not supported in that the AUD- and PTS-only groups generally did not differ from each other, and the PTS-only group did not differ from controls. These findings will be further discussed in turn.

Study findings showed that the comorbid group demonstrated greater endorsement of EA compared to the control group and PTS-only group, although the other groups did not differ. Although this may indicate greater severity found in a comorbid population, this may also suggest EA as a possible transdiagnostic risk factor for comorbid PTSD-AUD, particularly in the context of trauma exposure. Present study findings, wherein the PTS-only group did differ from the control group, align with prior EA work, which found consistent, negative associations between DT and PTSD symptom severity (Table 1; e.g., Feingold & Zerach, 2021; Khan et al., 2018; Paltell & Berenz, 2022; Zegel et al., 2019). It is important to note that much PTSD work to date has neglected to assess substance use or alcohol use; thus, these results provide important novel findings of how EA impacts these comorbid associations.

Results for the models examining DM subscales are more nuanced. The comorbid and AUD-only groups endorsed greater DM across all three subscales (i.e., social, enhancement, coping) than the control group (in line with our hypothesis). Both groups were also greater than the PTS-only group in the social subscale, and the comorbid group was greater than the PTS-only group in the coping subscale (Table 1). These findings are consistent with the literature examining military populations with current AUD, demonstrating that greater drinking motives are associated with greater alcohol problems (Bresin & Mekawi, 2021; Cheng et al., 2017). The coping subscale finding for the comorbid group further supports the specific relevance of coping motives to PTSD-AUD comorbidity (e.g., coping motives mediating the relationship between PTSD and harmful alcohol use; Cheng et al., 2017; Hammarberg et al., 2017; Luciano et al., 2022; Miller et al., 2017). The association of social motives with both comorbid and AUD-only groups highlights the relevance of this motive for increased AUD risk alone or in combination with PTSD. Although social motives are externally sourced motives and less associated with alcohol misuse, findings demonstrate their continued relevance for disordered alcohol use in a combat veteran sample (Cooper, 1994; Gonzalez et al., 2009) and when correlated with coping motives across multiple populations highlighting that the combination of high motives may be particularly risky (Kuntsche et al., 2005). This overlap of social and coping drinking motives in a military population associated with AUD and PTSD-AUD is important to consider, given the social and cultural norms of alcohol use within the military. Qualitative studies using a military personnel sample suggest alcohol use is viewed as an acceptable way to relax and cope with the stress of military life, and the ability to tolerate heavy alcohol use is viewed as a desirable trait in the military community (Besse et al., 2018; Kiernan et al., 2018). As a result, alcohol misuse can be normalized to the point of functional impairment in military populations (Jones & Fear, 2011), making both social and coping motives prevalent, risky, and a shared risk factor for the high rates of comorbid PTSD-AUD found in veterans and current military personnel.

There are notable commonalities among the proposed risk categories that warrant attention. EA is described in the literature as the use of various coping methods to modify the frequency, duration, or form of unwanted thoughts, feelings, and similar experiences. Individuals exhibiting EA may consistently use certain coping skills (e.g., denial) regardless of the situation’s demands or the strategy’s effectiveness. On the other hand, coping is a specific strategy aimed at altering one’s internal experience. Although coping can include broader avoidance behaviors that might be considered EA (e.g., mental disengagement, denial), these elements are distinct. Several studies have explored the relationship between these two constructs and found that, while they share similarities, they are separate concepts that should be distinguished (Karekla & Panayiotou, 2011; Kashdan et al., 2006). Additionally, individuals with low distress tolerance often struggle with managing negative emotions and are more likely to use maladaptive coping strategies, such as experiential avoidance, to escape their discomfort (Larrazabal et al., 2022). Yet, distress tolerance difficulties are also associated with other coping strategies.

The comorbidity between PTSD and AUD is a complicated relationship that is highly clinically impactful. Identifying constructs that link these two disorders can help inform treatment to identify potentially modifiable targets that underlie the relationship, leading to high comorbidity rates. For instance, some interventions directly address and attempt to modify behaviors similar to these constructs (e.g., Acceptance and Commitment Therapy [ACT] directly addresses EA during treatment procedures and process-based therapy targets key mediators and moderators of disorders) (Hofmann & Hayes, 2019; Zakiei et al., 2021). Further examination of the role of these constructs in the comorbid PTSD-AUD relationship is warranted to support focused treatment approaches. Further, addressing possible deficits (e.g., EA, DT, and DM domains), or shared risk and resilience factors, before treatment is an often-cited approach to improve treatment outcomes for this comorbidity (Sloan et al., 2017).

There are several limitations to be noted in this current study. First, the modest sample size for clinical groups could hamper our ability to distinguish smaller associations and group differences. We also note that a follow-up multivariate examination of the pattern of group ranks demonstrated on study measures (i.e., the comorbid group was always “worst,” the control group was always “best,” and the AUD group was “second worst” on most constructs) has a probability of < .05. This suggests that there were likely other meaningful, but small, differences that were not detected in this sample. Further replication, in larger, more diverse samples, as well as utilizing other statistical approaches suited to smaller samples to answer these broader questions of shared risk, is needed. Second, this study was cross-sectional, examining current diagnostic status, and thus cannot inform upon disorder onset or establish a causal association between the constructs of interest and development of PTSD and/or AUD. Future work utilizing a longitudinal format would allow for examining directionality across variables and the potential for mediation effects. Third, although we modeled constructs separately given the aims of this work, and supported by theoretical and statistical information from construct correlations, we note that future work could capitalize on their interrelationships using other statistical methods, such as latent factors. Lastly, the present sample was drawn from a parent study using a white male combat veteran population from the Iraq and Afghanistan conflicts between the ages of 21–45 years. This sample homogeneity limits the generalizability of these results. Further work is needed to examine the relevance of these constructs for PTSD-AUD comorbidity across various races, ages, gender, and trauma types. Finally, due to the aims of the parent study, current depressive episode and other substance use disorders were exclusionary. These conditions are also highly comorbid with PTSD and/or AUD, and future work examining factors relevant to their comorbidity would be useful.

Although the pattern of findings from the present study provides some support for shared factors underlying the relationship between PTSD and AUD, results also suggest that these relationships are quite nuanced. We found greater severity with comorbidity across the psychological constructs, demonstrating the need for more work that includes a comorbid population explicitly. Investigating other potentially relevant constructs would be beneficial to understand the etiology of this relationship and further inform treatment for comorbid PTSD-AUD utilizing these modifiable constructs.

Funding Statement

This work was supported by K01 [award No. AA025692] from the National Institute on Alcohol Abuse and Alcoholism (NIAAA; PI: Sheerin). This work was in part supported by CTSA [award No. UL1TR002649] from the National Center for Advancing Translational Sciences (NCATS). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIAAA or the NCATS or the National Institutes of Health. Dr. Amstadter’s time is partially funded by NIAAA [grants K02 AA023239 and R01 AA020179].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

These study data come from a larger ongoing study which will be made publicly available at the completion of that parent study, but materials and analysis code for this study are available by emailing the corresponding author.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08995605.2024.2387914

Open scholarship

This article has earned the Center for Open Science badge for Preregistered. The materials are openly accessible at https://osf.io/6vj5a.