Effects of Oxytocin on Stress Reactivity and Craving in Veterans with Co-Occurring PTSD and Alcohol Use Disorder

Abstract

Posttraumatic stress disorder (PTSD) and alcohol use disorder (AUD) are highly prevalent and commonly co-occur. The dual diagnosis of PTSD/AUD is associated with serious negative sequalae and there are currently no effective pharmacological treatments for this comorbidity. Both PTSD and AUD are characterized by dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which helps modulate stress reactivity. Oxytocin, a neuropeptide that attenuates HPA axis dysregulation, may be beneficial for individuals with co-occurring PTSD/AUD. Thus, the current study examined the effects of intranasal oxytocin (40 IU) as compared to placebo on stress reactivity (e.g., cortisol) as well as subjective alcohol craving in response to a laboratory stress task (Trier Social Stress Task). Participants were 67 male U.S. military veterans with current PTSD and AUD (oxytocin n=32, placebo n=35; overall mean Age=49.06 years). Baseline cortisol levels were examined as a moderator of outcome. The findings revealed that oxytocin marginally attenuated cortisol reactivity in response to the stress task. Furthermore, oxytocin’s effect was moderated by baseline cortisol level, such that oxytocin mitigated cortisol reactivity to a greater extent among participants with higher, as compared to lower, baseline cortisol. Oxytocin did not reduce craving. Although preliminary, the findings are the first to examine oxytocin in co-occurring PTSD/AUD. The findings from this study contribute to the growing literature examining the potential utility of oxytocin among individuals with psychiatric disorders, such as PTSD and substance use disorders. NCT02058251. Keywords: PTSD; alcohol; craving; comorbidity; oxytocin; social stress

Introduction

In comparison to the general population, military veterans are at increased risk of developing both posttraumatic stress disorder (PTSD) and alcohol use disorders (AUD) (Petrakis, Rosenheck, & Desai, 2011; Seal et al., 2011). PTSD and AUD frequently co-occur, and some of the highest rates of co-occurrence (e.g., approximately 40%) have been found among veterans (Petrakis et al., 2011; Pietrzak, Goldstein, Southwick, & Grant, 2011). Individuals with co-occurring PTSD/AUD commonly experience other psychiatric problems such as depression, anxiety, and suicidality, as well as poorer treatment outcomes, longer duration of substance use, and more treatment episodes (Flanagan, Korte, Killeen, & Back, 2016). Thus, developing effective treatments for this debilitating dual condition is a public health priority. Few studies to date have examined pharmacotherapies for co-occurring PTSD/AUD and those that have suggest only a modest response (Taylor, Petrakis, & Ralevski, 2017).

Oxytocin is a hypothalamic neuropeptide with emerging evidence supporting its role in the treatment of PTSD. Preclinical studies demonstrate oxytocin’s anxiolytic and anti-stress properties (Eskandarian et al., 2013; Missig, Ayers, Schulkin, & Rosen, 2010), and clinical studies show that oxytocin improves social cognition and behaviors (MacDonald & MacDonald, 2010). Preclinical and clinical models also suggest that oxytocin can ameliorate fear responses (Labuschagne et al., 2010; Petrovic, Kalisch, Singer, & Dolan, 2008) and enhance fear extinction in healthy individuals (Acheson et al., 2013). However, various individual and contextual factors such as sex, social context, and psychiatric history may influence human oxytocin response (Bartz, Zaki, Bolger, & Ochsner, 2011; Flanagan, Baker, McRae, Brady, & Moran-Santa Maria, 2015; Flanagan et al., 2018).

The literature also suggests that oxytocin may be a promising candidate in the treatment of substance use disorders (see McGregor & Bowen, 2012, for review). Oxytocin has been shown to reduce withdrawal symptoms, craving, and self-administration among individuals with dependence on drugs such as marijuana and stimulants (Carson et al., 2010; McRae-Clark, Baker, Moran-Santa Maria, & Brady, 2013). However, the literature is currently limited by a scarcity of studies examining AUD as compared to drug use disorders. Oxytocin has been shown to block the development of alcohol tolerance and ameliorate alcohol withdrawal-related central nervous system hyperexcitability in preclinical research (Szabo, Kovacs, Szekeli, & Telegdy, 1985). A more recent study found that oxytocin significantly reduced alcohol withdrawal symptoms among treatment-seeking participants (Pedersen et al., 2013). Other research suggests that oxytocin may reduce the effects of social stress on alcohol craving and consumption (Mitchell, Arcuni, Weinstein, & Woolley, 2016; Peters, Slattery, Flor, Neumann, & Reber, 2013). There is also growing preclinical evidence that oxytocin reduces alcohol self-administration and reinstatement of alcohol seeking in mice (King, Griffin, et al., 2017; King, McGinty, & Becker, 2017).

Previous studies suggest that oxytocin’s behavioral outcomes are attributable, in part, to its ability to attenuate hypothalamic-pituitary-adrenal axis (HPA) dysregulation (Cardoso, Kingdon, & Ellenbogen, 2013). Cortisol is among the most commonly examined markers of HPA axis function in human oxytocin studies (see Cardoso et al., 2014, for review). While the specific mechanisms by which oxytocin might influence human HPA axis function, and cortisol reactivity specifically, is currently a topic of debate in the clinical literature, cortisol reactivity is associated with both PTSD and AUD. For example, HPA axis dysregulation (commonly measured by salivary cortisol) is associated with greater alcohol and drug craving, withdrawal, and relapse in both preclinical and clinical studies and may reflect increased risk for developing a substance use disorder (Breese, Sinha, & Heilig, 2011; Lovallo, 2006). Similarly, HPA axis dysregulation, particularly in the form of heightened cortisol reactivity, is a well-established correlate of PTSD diagnosis among civilians and veterans (Morris, Hellman, Abelson, & Rao, 2016; Steudte- Schmiedgen et al., 2015).

The scarcity of studies examining the effects of oxytocin among individuals with co-occurring PTSD and AUD is an important gap in the literature. The current study is a preliminary effort to bridge that gap by examining the effects of a single dose of intranasal oxytocin on stress reactivity (i.e., cortisol response) and craving in response to laboratory-induced stress. We hypothesized that participants randomized to oxytocin, as compared to placebo, would demonstrate more attenuated cortisol reactivity and alcohol craving in response to the task. Because emerging hypotheses suggest that oxytocin may attenuate stress reactivity more effectively among individuals with less adaptive stress responses compared to healthy individuals (Cardoso, Linnen, Joober, & Ellenbogen, 2012; Flanagan et al., 2015) we also examined the moderating role of baseline cortisol levels on oxytocin’s response.

Methods

Participants

All study procedures were IRB-approved (Medical University of South Carolina IRB-I PRO#34567; Study Title: Oxytocin Suppresses Substance Use Disorders Associated with Chronic Stress). Participants were 67 male U.S. military veterans. Participants were recruited primarily via advertisements in the local newspapers and online (e.g., Craigslist), flyers posted in relevant clinics throughout the Ralph H. Johnson VA hospital, and through the Substance Treatment and Recovery (STAR) clinic as well as the PTSD Clinic Team (PCT) at the VA. Inclusion criteria were: aged 21–65 years old; meet DSM-5 diagnostic criteria for current PTSD and AUD (additional substance use disorders were allowed); sufficient intellectual functioning to complete study procedures as indicated by a Mini Mental Status Exam (Folstein, Folstein, & McHugh, 1975) score > 26; at least 5 days of abstinence from alcohol (self-reported) and other substances (except caffeine or nicotine; urine drug screen) prior to completing the laboratory testing; and stabilization of any psychotropic medications for at least four weeks prior to participation. Exclusion criteria included: history of or current psychotic disorders; currently taking psychotropic medications likely to reduce alcohol craving; suicidal or homicidal ideation and intent; major medical illness (e.g., endocrine, cardiovascular, pulmonary disease) or other unstable medical or psychiatric conditions that could affect study participation; significant alcohol withdrawal symptoms as evidenced by a Clinical Institute Withdrawal Assessment of Alcohol (Sullivan, Sykora, Schneiderman, Naranjo, & Sellers, 1989) score ≥ 8; and systolic blood pressure ≥ 160 mmHg, diastolic blood pressure ≥ 100 mmHg and resting pulse rate ≥ 100 bpm.

Approximately one-third (35.8%) of the sample was recruited from the substance abuse treatment clinic at the local VA while the remaining two-thirds of the sample consisted of veterans who learned about the study through resources in the community and on the internet. A total of 107 individuals completed baseline assessments and 73 (67 men, 6 women) met eligibility requirements and were enrolled in the study. Emerging literature indicates important sex differences in both neurobioloical and behavioral oxytocin outcomes (Ditzen et al., 2012; Flanagan et al., 2018; Hoge et al., 2014; MacDonald, 2015; Rilling et al., 2014; Yao et al., 2014). Additionally, the sample of women enrolled (3 oxytocin, 3 placebo) was insufficient to allow for appropriate statistical group comparisons. Thus, the present analyses focus on male participants (n = 67; 32 oxytocin, 35 placebo).

Measures

Diagnostic evaluation.

The Mini-International Neuropsychiatric Interview (MINI; Sheehan et al., 1998) was used to asses, as DSM-5 psychiatric diagnoses, including AUD. The Clinician Administered PTSD Scale (CAPS-5; Weathers, Blake, & Schnurr, 2013), a structured clinical interview with a scoring range of 0–80, assessed DSM-5 diagnostic criteria for PTSD. In addition, PTSD symptom severity was assessed using the 17-item, self-report PTSD Checklist-Military version (PCL-M; Weathers, Huska, & Keane, 1991). The PCL-M has a scoring range of 17–85. Cronbach’s α=.95.

Substance Use and Craving.

Breathalyzer tests were used to evaluate acute intoxication. Urine drug screens were administered using the On Track Test Cup ^®. Quantity and frequency of alcohol use (total drinking days; average number of standard drinks per drinking day) was assessed during the 60 days prior to participation using the Time Line Followback (TLFB; Sobell & Sobell, 1992), a calendar-assisted, semistructured interview. Alcohol use problem severity was assessed using the 10-item, self-report Alcohol Use Disorders Identification Test (AUDIT; Babor, Higgins-Biddle, Saunders, & Monteiro, 2001) with a range of 0–40. Cronbach’s α=.83 in this sample. Alcohol craving was measured using a modification of the Within Session Rating Scale (Childress, McLellan, & O’Brien, 1986). This 100 mm visual analogue scale is anchored from 0 = none to 10 = extreme.

Procedures

All procedures were IRB-approved and all participants completed informed consent prior to study involvement. Participants completed study procedures in a single 4-hour laboratory visit. Participants were scheduled at the same time of day (7:00 a.m.) to control for diurnal variations in HPA axis function. Following provision of a negative breathalyzer test and urine drug screen, participants completed a battery of self-report and interview measures. Smokers were asked to abstain from smoking during study procedures and offered a nicotine patch prior to participation. Nicotine patch dosage was determined by the number of cigarettes participants smoked daily (≥20 cigarettes/day=21mg; 10–19 cigarettes/day = 14 mg patch; 5–9 cigarettes/day =7 mg patch). Raw oxytocin powder was obtained from Professional Compounding Centers of America. Oxytocin and placebo were compounded using a standardized recipe by Investigational Drug Services at the Medical University of South Carolina, which also implemented the block-design randomization scheme. Medication was distributed by the pharmacy in identical nasal spray bottles. Following standardized procedures, participants were instructed on proper self-administration and observed by research staff. Participants were instructed to blow their nose, exhale through their nose, then spray into one nostril while inhaling, alternating nostrils until the 40 international units (IU) dose was achieved. The concentration of the nasal spray ranged from 40 IU/.8mL to 40 IU/1.2mL between batches. Participants self-administered 4–6 sprays in each nostril. Participants were randomly assigned in a doubleblind manner (1:1) to receive oxytocin (40 IU) or placebo.

Laboratory procedures are described in Table 1. At approximately 7:00 am, participants arrived at the office and completed a battery of self-report and interview assessments. At 8:30 am and 8:50 am, pre-task baseline saliva samples were collected to measure cortisol. The mean of these values was used as the baseline cortisol value, prior to administration of oxytocin or placebo (Time 1). Participants self-administered 40 International Units (IU) of intranasal oxytocin or placebo (i.e., saline) at approximately 9:15 am. Participants then completed a 45-minute rest period to allow the medication to take effect. The timing of medication administration is based on extensive past research in our group and others (Cardoso, Ellenbogen, Orlando, Bacon, & Joober, 2012; Flanagan et al., 2015; Flanagan, Sippel, Wahlquist, Moran- Santa Maria, & Back, 2017; Guastella et al., 2013; McRae-Clark et al., 2013; van Zuiden et al., 2016). During this time, saliva samples were collected at 9:30 and 9:45 (Times 2 and 3, respectively). The laboratory stress task, described below, began at approximately 10:00 am. Post-task saliva samples were collected immediately following the stress task (approximately 10:15; Time 4), and again at 5- (Time 5), 15- (Time 6), and 30- minutes (Time 7) after the task. Participants were then debriefed and compensated for their time.

Table 1.

Laboratory Testing Procedures

Time	Study Time Point	Procedure
7:00 am		Informed consent and assessment battery
8:30 am	Time 1	Pre-task baseline cortisol assessment #1
8:50 am		Pre-task baseline cortisol assessment #2
9:15 am		Medication self-administration (40 IU oxytocin or placebo)
9:16 am		Rest period begins
9:30 am	Time 2	Cortisol assessment #3
9:45 am	Time 3	Cortisol assessment #4
9:55		Rest period ends
10:00–10:15 am		Trier Social Stress Task (TSST)
10:15 am	Time 4	Post-task cortisol assessment #5 (immediately post-TSST)
10:20 am	Time 5	Post-task cortisol assessment #6 (5-minutes post-TSST)
10:45 am	Time 6	Post-task cortisol assessment #7 (30-minutes post-TSST)
11:15 am	Time 7	Post-task cortisol assessment #8 (60-minutes post-TSST)
11:15 am		Debriefing, compensation, and discharge

Note. IU=international units.

Social Stress Paradigm Procedures

The Trier Social Stress Task (TSST; Kirschbaum, Pirke, & Hellhammer, 1993), a standardized, 15- minute laboratory task was employed in the current study. During the TSST, participants were told they would perform a speech about their “dream job” to an audience and do arithmetic calculations. At 10:00 am, the experimenter gave participants 5 minutes to prepare the speech. At 10:05 am, three individuals unfamiliar to the participant (the audience) entered the room and were seated. The participant was then instructed by one audience member (the spokesperson) to stand and begin his/her prepared speech (without notes) for 5 minutes. At the end of the speech (10:10 am), the participant was instructed to conduct serial subtractions (13 from 1,022) as quickly and accurately as possible. The arithmetic calculations continued for 5 minutes. At the end, the spokesperson instructed the participant to stop and be seated, and the audience left the room. In order to ensure that the TSST evoked a stress reaction from participants, subjective reports of anxiety before and after the TSST were measured using a modified visual analog scale anchored from 0 = none to 10 = extreme (Childress et al., 1986).

Neuroendocrine assay

Unstimulated salivary samples were collected by passive drool in polypropylene vials and immediately iced. Samples were aliquoted into 1.8 nunc tubes and saliva was then frozen at-70 °C until ready for assay. Samples were assayed in duplicate using a high sensitivity salivary cortisol enzyme immunoassay system that has an intra-assay precision of 3.35%—3.65% with a lower sensitivity limit of <0.003 μg/dL (Salimetrics, LLC). Samples were analyzed using a PowerWave HT Microplate Spectrophotometer in conjunction with a Precision Series Automated Liquid Handling System (BioTek Instruments, Inc.).

Statistical Analyses

Area under the curve (AUC) values were used to capture overall change in cortisol levels and craving across all seven time points. AUC values were computed with respect to ground (as opposed to baseline cortisol) so that baseline cortisol levels could be examined as a moderator of the association between treatment condition and AUC post-baseline (Pruessner, Kirschbaum, Meinlschmid, & Hellhammer, 2003). Positive AUC values indicate an increase in cortisol or craving across all time points. Negative AUC values indicate a decrease in cortisol or craving levels across all time points. The individual time intervals (in minutes) between assessments were used when computing AUC; this approach is recommended when the intervals may not be constant.

Path analysis was used to examine the effect of drug condition (0 = oxytocin, 1 = placebo) and baseline cortisol levels on post-baseline AUC-calculated cortisol levels. A similar model examining craving was run separately. In addition to baseline cortisol levels, smoking status was included as a covariate as previous literature has demonstrated its effects on HPA axis function (Rohleder & Kirschbaum, 2006; Steptoe & Ussher, 2006). Subsequently, baseline cortisol levels were centered and a baseline cortisol (+/− 1 standard deviation [SD]) by drug condition interaction was examined in relation to cortisol reactivity and craving, also controlling for smoking status.

Results

Descriptive Statistics

Demographic and clinical characteristics are presented in Table 2. Participants had a mean CAPS score of 35.2 (SD=9.1), and a PCL-M score of 45.2 (17.8). On average, participants reported 4.41 drinking days and 1.03 drinks per drinking day on the TLFB. Despite the low frequency and quantity of recent alcohol use at baseline, the mean AUDIT score was 23.50 (SD=8.80) and the majority of the sample (n=65, 97.01%) met diagnostic criteria for severe AUD. The largest proportion of participants in this sample served in the Army (n=36; 53.7%). Participants had served an average of 5.64 years (SD=4.17 years) and 13 participants (20.9%) served in the conflicts in Iraq and/or Afghanistan. Notably, similar proportions (53.1% of participants in the oxytocin condition and 40.0% of participants in the placebo condition) correctly identified their drug condition, suggesting integrity of the double-blind design.

Table 2.

Sample Demographic and Clinical Characteristics

	Total Sample N = 67	Placebo Group n = 35	Oxytocin Group n = 32
	Mean (SD) or n (%)			p
Demographics
Age	49.06 (10.41)	48.09 (11.41)	50.13 (9.26)	.43
Education	13.52 (1.81)	13.69 (1.53)	13.34 (2.09)	.45
Race				.004
African American/Black	41 (61.2)	16 (45.7)	25 (78.1)
Caucasian/White	23 (34.3)	18 (51.4)	5 (15.6)
Other	2 (3.0)	1 (2.9)	2 (6.3)
Employment				.24
Unemployed	46 (68.7)	26 (74.3)	20 (62.5)
Employed Full-Time	7 (10.4)	3 (8.6)	4 (12.5)
Disabled/Retired/Student	13 (19.4)	6 (17.2)	7 (21.9)
Hispanic ethnicity	2 (3.0)	2 (5.9)	0 (0.0)	.49
Branch of Military Service				.79
Army	36 (53.7)	19 (54.3)	17 (53.1)
Navy	10 (14.9)	4 (11.4)	6 (18.8)
Marine Corps	8 (11.9)	4 (11.4)	4 (12.5)
Air Force	8 (11.9)	5 (14.3)	3 (9.4)
Other	5 (7.5)	3 (8.6)	2 (6.3)
Clinical characteristics
CAPS Total Score	35.08 (9.14)	35.77 (10.22)	34.31 (7.88)	.52
PCL-M Total Score	45.22 (17.83)	45.89 (17.63)	44.50 (18.30)	.75
TLFB-Drinking Days	4.41 (7.16)	4.85 (7.88)	3.94 (6.40)	.61
TLFB-Amount per drinking day	1.04 (2.03)	.91 (1.94)	1.17 (2.15)	.62
AUDIT	23.50 (8.80)	23.34 (8.90)	23.68 (8.83)	.88
SUD Diagnosis	378 (56.7%)	21 (60.0%)	17 (53.1%)	.58
Severe AUD diagnosis	65 (97.0%)	35 (100%)	30 (93.8%)	.50
Prior/Current AUD Treatment	58 (86.8%)	32 (91.4%)	26 (81.2%)	.41
Prior/Current PTSD Treatment	37 (55.2%)	22 (62.9%)	15 (46.9%)	.40
Smoker	42 (62.7)	23 (65.7)	19 (59.4)	.59
Cortisol (ng/ml)	0.16 (.09)	0.17 (.09)	0.16 (.09)	.97

Note: Group differences in continuous characteristics were assessed using independent samples t-tests and differences in categorical characteristics were assessed using normal chi-square tests of independence or Fisher’s exact tests. ng/ml = nanogram/millimeter. Ethnicity was missing for one participant. AUDIT=Alcohol use disorders identification test. CAPS= Clinician Administered PTSD Scale. PCL-M=PTSD Checklist Military version. TLFB= Time Line Followback. AUD= Alcohol Use Disorder. Time frame for TLFB assessment was past 60 days.

Repeated measures ANOVAs were conducted to validate that the stressor task evoked a stress response from participants. These analyses considered changes in subjective reports of anxiety and craving as well as changes in cortisol from time point 3 to time point 4. Subjective reports of anxiety and craving increased significantly (F = 29.71,p < .001 and F = 9.77,p < .05, respectively). Cortisol also increased but the increase did not reach significance (F = 1.47, p = .23).

Effects of Oxytocin Administration

Mean cortisol levels in each drug condition throughout the study are presented in Figure 1. Results indicate that there was a marginal effect of drug condition on cortisol reactivity (B = 3.52, p = .07),. As shown in Figure 2, veterans who received oxytocin, as compared to placebo, evidenced less cortisol reactivity in response to the stress task. Results of the path analytic model including drug condition, baseline cortisol, their interaction, and smoking status predicting cortisol reactivity are provided in Table 3. Results indicate that the association between drug condition and cortisol reactivity was qualified by a significant drug condition by baseline cortisol interaction (B = 38.84, p = .03). Probing the interaction revealed no significant group (oxytocin vs. placebo) differences at baseline cortisol levels one SD below the mean (B = .17, p = .99) but significant group differences at one SD above the mean (B = 6.67, p < .01). As shown in Figure 2, individuals with high baseline cortisol were significantly more likely than individuals with low baseline cortisol to demonstrate a decrease in cortisol production across time points if they received oxytocin. In contrast, individuals with high baseline cortisol were more likely to demonstrate an increase in cortisol production across time points if they received placebo. Drug condition was not associated with significant changes in stress-induced craving (B = .14, p = .25).

Figure 1.

Mean cortisol values across time points based on drug condition (oxytocin vs. placebo). Intranasal administration of oxytocin and placebo sprays occurred 25 minutes post-baseline. TSST = Trier Social Stress Task.

Figure 2.

Mean cortisol values across all time points based on drug condition (oxytocin vs. placebo) and baseline (BL) cortisol values (low = −1SD, high = +1SD). Intranasal administration of oxytocin and placebo sprays occurred 25 minutes post-baseline. TSST = Trier Social Stress Task. BLCort = Baseline cortisol level.

Table 3.

Effects of Oxytocin on Cortisol Reactivity as Moderated by Baseline Cortisol Levels

	B	SE	p	R2 (%)
Main Effects Model				30.4
Drug Condition	3.52	2.04	.08
Smoker	-2.35	2.10	.26
Moderation Model				44.3
Drug Condition	3.42	1.57	.03
Baseline Cortisol	0.09	29.33	1.00
Drug Condition X Baseline Cortisol	38.84	18.36	.03
Smoker	-1.61	1.65	.33

Note. SE = Standard error. Condition (0 = oxytocin, 1 = placebo). Smoker (0 = not smoker, 1 = smoker).

Given differences in racial composition in the oxytocin and placebo conditions, all models were reexamined including race (0 = Caucasian/White, 1 = other). The effect of drug condition on cortisol was marginally significant (B = 3.25 p = .06). There were no other differences in the results and race was not a significant covariate. Although there were not enough women enrolled in the study to examine gender differences in these effects, we did re-examine the path analytic model above including the 6 female participants. The effect of the drug condition on cortisol reactivity was attenuated (B = 2.58, p = ..0) as was the drug condition by baseline cortisol interaction (B = 19.04, p = .27).

Discussion

The present study examined the effects of a single administration of oxytocin on stress reactivity and craving in response to laboratory stress task among veterans with PTSD and AUD. The literature examining the effects of oxytocin among individuals with PTSD is growing rapidly, as is a separate body of research examining the effects of oxytocin among individuals with substance use disorders (see Koch et al., 2014; McGregor & Bowen, 2012; Olff et al., 2013; for review). The current study addresses an important gap in this growing literature by examining oxytocin among individuals with co-occurring PTSD and AUD. In addition, the current study adds to the existing literature by focusing on military veterans. To date, most of the clinical studies examining oxytocin have focused on civilians, and research demonstrates that veterans are disproportionately at risk for developing both PTSD and AUD (Petrakis et al., 2011). Examination of the effects of oxytocin, as compared to placebo, in response to a standardized laboratory stress task (Trier Social Stress Task) revealed that oxytocin was associated with a decrease in cortisol reactivity, especially among male veterans with high baseline cortisol levels. Examination of subjective craving revealed no group differences.

The current study adds to a growing literature suggesting that oxytocin’s neurobiological and behavioral effects are highly nuanced. When we consider the extensive literature linking heightened cortisol reactivity and long-term cortisol levels to numerous negative health risks and outcomes (Ennis et al., 2017; Wells et al., 2014), the current findings lend tentative support to the emerging hypothesis that oxytocin might be more effective among individuals with more salient health or interpersonal vulnerabilities. Previous oxytocin studies have identified various factors such as greater trauma exposure, certain psychiatric diagnoses, and the social context in which oxytocin is employed as important considerations when evaluating its clinical effects (Cardoso et al., 2014; Flanagan et al., 2015). The current finding that oxytocin more effectively mitigated cortisol reactivity among those participants who had higher baseline cortisol levels adds to this literature.

Additional literature regarding the complexity of cortisol reactivity must also be considered when interpreting the current findings. For example, studies suggest that blunted cortisol reactivity is also a marker of various health problems (Phillips, Ginty, & Hughes, 2013). With regard to mental health specifically, a recent review by Staufenbiel and colleagues (2013) emphasized the complicated association between cortisol reactivity and specific psychiatric diagnoses. Indeed, blunted cortisol reactivity has been linked with both PTSD and AUD in separate literatures. With regard to PTSD, studies have found associations between lower hair cortisol levels and PTSD diagnostic risk and symptom severity (Mouthaan et al., 2014; Steudte-Schmiedgen et al., 2015). However, a recent review found no consistent association between PTSD and cortisol levels, although this finding may be due to methodological inconsistencies in the measurement of cortisol across studies (Morris et al., 2016). Similarly, recent studies indicate that heavy alcohol consumption, particularly over the long-term, is associated with lower diurnal cortisol as well as lower cortisol reactivity to laboratory stress and alcohol cue tasks (Blaine, Seo, & Sinha, 2017; King, Hasin, O’Connor, McNamara, & Cao, 2016). One important limitation in the present study is the lack of a healthy control group. The current findings suggest that in a population of individuals with a dual diagnosis that confers a compounded risk for a blunted cortisol stress response, those individuals presenting with the highest baseline cortisol levels appear to derive maximum benefit from oxytocin administration. Incorporating a healthy control group is an important next step for future studies as is enrolling a sample capable of thoroughly examining sex differences. In addition, considering the fact that race emerged as a factor that influenced the outcomes examined here, future studies would benefit from recording and examining both participant and TSST confederate race and ethnicity. Some studies suggest that race, perceived difference in social status, and implicit racial bias has the ability to influence social stress reactivity (Gevonden et al., 2016; Mendes, Gray, Mendoza-Denton, Major, & Epel, 2007; Page-Gould, Mendes, & Major, 2010). Thus, it is possible that in this sample, social stress reactivity might be influenced by racial similarity or difference between participants and the TSST audience.

While much of the existing clinical research on oxytocin has focused on drug use disorders (e.g., cocaine, marijuana), the current study focused on AUD. Previous studies suggest that oxytocin may effectively treat factors that commonly underlie many substance use and psychiatric disorders, such as depleted social reward and dysregulation in corticolimbic brain regions (McGregor & Bowen, 2012). A recent preclinical study found that oxytocin reduced alcohol consumption in mice (King, Griffin, et al., 2017). Pederson and colleagues (2013) found that oxytocin successfully mitigated alcohol withdrawal symptoms in treatment-seeking individuals. Another recent study however found that oxytocin was associated with greater alcohol craving but reduced appetitive approach behaviors among a small sample of individuals (N=32) with AUD (Mitchell et al., 2016). In the current study, group differences in craving following the stress task were not observed. Overall, alcohol craving was low throughout the laboratory visit for both groups, which may have influenced the findings. While the TSST significantly influenced cortisol reactivity among individuals in the placebo condition, it did not result in increased alcohol craving. This finding could be related, in part, to several factors. First, many of the participants were treatment- seeking and recruited from the substance use treatment clinic at the local VA hospital. Despite meeting diagnostic criteria for a current AUD, some had already taken steps to significantly reduce their alcohol use, which may have reduced their craving as evidenced by low levels of alcohol use at baseline. An additional methodological consideration is that the study took place early in the morning in order to control for diurnal variation in cortisol, which may have influenced the outcomes examined here. Future research would benefit from testing the effects of oxytocin at different times of the day and including an alcohol cue paradigm to prime craving. It is also possible that a single administration of oxytocin or the dose employed (40 IU) in the current study are insufficient to modify craving among veterans with PTSD/AUD. Further research is needed to help address these important clinical questions.

Although preliminary, the current study contributes to a growing literature aimed at ultimately developing effective pharmacological treatments for co-occurring PTSD and AUD. Numerous medications have been investigated and there are some promising findings but also substantial room for improvement (Flanagan et al., 2016; Taylor et al., 2017). The persistence of AUD symptoms is among the most commonly cited barriers to maximizing treatment efficacy for co-occurring PTSD/AUD (Badour et al., 2017). The effects of oxytocin on HPA axis stress reactivity in the current study suggest that oxytocin may hold promise and should be further evaluated among individuals with PTSD/AUD. Future studies should be adequately powered to examine additional modifiers of oxytocin response, such as PTSD and AUD symptom severity, in order to identify subgroups of individuals for whom oxytocin might be most beneficial.

Conclusions

Co-occurring PTSD and AUD is a common and detrimental dual diagnoses with limited effective treatment options currently available. Separate literatures have established that oxytocin has the potential to reduce symptoms of both disorders. The current study is the first to examine oxytocin among individuals with comorbid PTSD/SUD. This study is methodologically and theoretically informative for future oxytocin studies in the area of co-occurring PTSD/AUD. The present findings provide modest support to the growing evidence suggesting that oxytocin may attenuate cortisol reactivity in select clinical populations. The results also add to the literature suggesting that the adaptive neurobiological effects of oxytocin may not directly translate to hypothesized behavioral effects, and identifying ways to translate the neurobiological effects of oxytocin into meaningful clinical applications is an important remaining gap in the literature.

Public Significance Statements.

Effective pharmacological interventions for co-occurring posttraumatic stress disorder (PTSD) and alcohol use disorders (AUD) are scant. This study examined the effects of oxytocin versus placebo on social stress and associated cortisol reactivity among Veterans with co-occurring PTSD and AUD. Results from this study provide modest support to existing literature suggesting that oxytocin may attenuate cortisol reactivity in select clinical populations.