Sex Differences across Retrospective Transitions in Posttraumatic Stress and Substance Use Disorders

Abstract

Objectives:

Concurrent substance use disorder (SUD) and posttraumatic stress disorder (PTSD) occur at high rates and are typically associated with poor treatment outcomes in both sexes. However, women have a propensity to cope with increased negative affect via substance use in comparison to men; thus, it is important to elucidate the sex-specific bidirectional relationships between SUD and PTSD to improve our understanding of concurrent SUD/PTSD in men and women.

Methods:

Using data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-Wave 3; n = 36,309), the present study evaluated the impact of sex on the relationship between past-year SUDs (new, remitted, ongoing), including alcohol and drug use, and retrospective transitions in new vs. absent and ongoing vs. remitted diagnoses of PTSD. Additionally, the impact of sex was explored in models examining past year PTSD (new, remitted, ongoing) and retrospective transitions in new vs. absent and ongoing vs. remitted diagnosis of SUDs. Diagnostic transitions were based on retrospective reporting.

Results:

Results indicated that new, remitted, and ongoing SUDs increase the likelihood of new PTSD diagnoses (OR range = 2.53–8.11; p < 0.05). Among individuals with ongoing drug use disorders (DUD), there were greater odds of ongoing PTSD (OR = 2.10, p < 0.01). When examining the relationship reciprocally, new, remitted, and ongoing PTSD increased the likelihood of new SUDs (OR range = 2.50–8.22; p < 0.05), and ongoing PTSD increased the likelihood of ongoing SUD and DUD (OR = 1.40, 1.70, respectively; p < 0.05). Sex-specific analyses revealed that the relationship between PTSD and SUDs varies between sexes, particularly among women. For instance, women with new PTSD had higher odds of SUDs, and women with ongoing PTSD were almost 2.5 times more likely to have an ongoing DUD. Women with a new PTSD diagnosis were more likely to be diagnosed with a new SUD (OR = 3.27) and an ongoing DUD (OR = 3.08).

Conclusions:

Results indicate a bidirectional relationship between PTSD and SUD that is in many instances larger in women. Thus, illustrating potential sex-specific differences in underlying mechanisms implicated in SUD/PTSD, warranting additional research.

Substance use disorders (SUDs) are highly prevalent in the United States; ~10% of the population has met diagnostic criteria for a drug use disorder (DUD) during one’s lifetime, while 3.9% of the population has met such criteria in the past twelve months (Grant et al., 2016). Similarly, high rates of alcohol use disorders (AUDs) are observed both over one’s lifetime and within the past twelve months, with 29.1 and 12.9% of the population meeting diagnostic criteria for AUD, respectively (Grant et al., 2015).

The majority of adults diagnosed with substance use disorders have been exposed to a traumatic event, with estimates that 95% of individuals with a SUD are exposed to trauma (Vujanovic & Back, 2019). Such high rates of trauma exposure unsurprisingly result in high rates of concurrent PTSD among individuals with SUDs. In fact, estimates indicate that 26–60% of individuals with a SUD also meet the criteria for PTSD (Vujanovic & Back, 2019).

Conversely, individuals with PTSD are at heightened risk to also meet diagnostic criteria for a DUD, both within the past year and across one’s lifetime (OR’s = 1.6 and 1.5, respectively), as well-being more vulnerable to an AUD over one’s lifetime (OR = 1.3; Grant et al., 2015, 2016). These relationships are especially strong among women, who despite being less likely than men to report experiencing a potentially traumatic event, are twice as likely to meet diagnostic criteria for PTSD (Tolin & Foa, 2006). Such increased rates of concurrent substance use disorder and PTSD, place both men and women at increased susceptibility, as these diagnoses often lead to more complex diagnostic presentations, including more severe trauma-related symptomatology, and poorer treatment outcomes (Bedard-Gilligan et al., 2018; Blanco et al., 2013; Flanagan et al., 2016; Saladin et al., 1995).

There are several pathways for the development of concurrent substance use disorders and PTSD (Berenz et al., 2019; Jacobsen et al., 2001). For example, individuals initially diagnosed with a SUD may be more likely to be vulnerable to the exposure of traumatic events and have an increased vulnerability for developing subsequent PTSD (Berenz et al., 2019; Jacobsen et al., 2001). Among a sample of college students, alcohol-related sexual assaults were associated with increased alcohol use (Kaysen et al., 2006) and substance use mediated the relationship between PTSD symptoms and re-victimization among sexual assault survivors in a similar sample (Messman-Moore et al., 2009).

Additionally, another example suggests that individuals use substances to escape negative affect, and trauma-related symptomatology (e.g., arousal symptoms, such as hypervigilance and irritability; Berenz et al., 2019; Jacobsen et al., 2001). PTSD and substance use may synergistically increase symptomatology, as ongoing substance use leads to a period of withdrawal, characterized by heightened negative affect and stress, which in turn leads to heightened PTSD symptoms (Berenz et al., 2019; Koob & Volkow, 2016). For example, ~20% of individuals with PTSD report using substances to manage trauma-related symptomatology (Leeies et al., 2010) and in a longitudinal study of National Guard troops, new AUD was preceded by PTSD following deployment (Kline et al., 2014). Thus, there is a bidirectional relationship regarding the development of concurrent substance use and PTSD, in which both PTSD and substance use disorders may increase the vulnerability and susceptibility to develop the other disorder simultaneously (Berenz et al., 2019; Jacobsen et al., 2001).

These relationships between substance use disorders and PTSD are especially important among women, who are more likely to utilize substances to regulate negative affect and stress (Becker & Koob, 2016; Koob & White, 2017; Peltier et al., 2019). Women are more than twice as likely as men to be diagnosed with comorbid PTSD and alcohol use (Sonne et al., 2003) and women report higher rates of drinking to cope compared to men (Kachadourian et al., 2014; Lehavot et al., 2014). Trauma exposure has been more closely associated with hazardous drinking in women compared to men (Kachadourian et al., 2014). Additionally, a clinical study of men and women with concurrent PTSD and AUD found that motivation to drink to cope was associated with both sexes, but drinking to cope was associated with average alcohol use for women whereas drinking for enhancement motives was associated with men (Lehavot et al., 2014).

It is important to note that there are also sex differences in trauma exposure and PTSD symptoms; men are more likely to experience traumatic events, but women are more likely to develop subsequent PTSD from trauma exposure (Torchalla & Nosen, 2019). In light of this evidence, it is likely that varying trauma exposures contribute to this sex difference (Torchalla & Nosen, 2019). A review by Tolin and Foa (2006) illustrated that women are more likely to endorse adult and childhood sexual abuse, while men are more likely to endorse physical assaults and combat exposure. Specifically, when examining a PTSD/substance use disorder population, over 70% of women endorsed a previous sexual assault (Back et al., 2003). Such differences in trauma type may account for heightened PTSD symptoms, as one study found that when controlling for trauma type (e.g., sexual trauma), sex differences in symptom severity became smaller (Guina et al., 2019).

Despite these sex differences, scant literature to date has sought to establish the role of sex in the development of PTSD among individuals with concurrent substance use disorders. The present study aimed to explore the bidirectional relationship of retrospective transitions SUD and PTSD diagnoses, as well as how such relationships vary between men and women with PTSD. Specifically, it was hypothesized that individuals with SUDs, including DUDs and AUDs, would show higher rates of new and ongoing PTSD diagnoses compared to those without SUDs. Additionally, we hypothesized that individuals with PTSD would be more likely to have new and ongoing SUDs. We predicted that these associations would be stronger in women compared to men.

Methods

Data source

The present study utilized data from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), Wave 3 (2012–2013). NESARC-III is a nationally representative, the cross-sectional dataset of information regarding the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5) substance use disorders and other psychiatric disorders, sponsored by the National Institute of Alcohol Abuse and Alcoholism (NIAAA). NESARC-III consists of 36,309 non-institutionalized men and women 18 years or older living in the United States. Each participant provided informed consent and completed an in-person, computer-assisted interview using the NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule (AUDADIS-5). Additional information regarding the sampling and methodology of NESARC-III is available elsewhere (Grant et al., 2014, 2015).

Substance use and psychiatric disorders

Relevant NESARC-III variables were coded into the following categories based upon substance use or PTSD symptoms over the past 12 months: absent, no diagnosis in the past year and no diagnosis before the past year (including any year before the past year); new, diagnosis in the past year but no diagnosis before the past year; remit, no diagnosis in the past year, but diagnosis before the past year; and ongoing, diagnosis in the past year and diagnosis before the past year. While the dataset is cross-sectional in nature, these categories indicated if there had been within-subject changes in diagnoses in the past year (e.g., within the past 12 months) or before the past year (e.g., <12 months ago). Thus, allowing for the assessment of current and prior diagnoses within the same interview. Variables include NESARC computed AUD and PTSD diagnoses. Of note, Criterion A traumas for PTSD diagnoses were assessed by querying participants across nineteen various “stressful life experiences” (e.g., serious injury/illness, natural disasters, physical/sexual abuse, assault, combat). Additionally, NESARC variables including cannabis use disorder, sedative use disorder, opioid use disorder, cocaine use disorder, stimulant use disorder, hallucinogen use disorder, inhalant/solvent use disorder, club drug use disorder, heroin use disorder, and other drug use disorders, were combined to create an overarching drug use disorder (DUD) variable. Participants who endorsed any of the above diagnoses were coded as having a DUD diagnosis. Additionally, AUD and DUD variables were combined (i.e., meeting criteria for AUD and/or DUD) to create an overall substance use disorder (SUD) variable.

Statistics

Demographic data were analyzed using SPSS, version 26 (IBM SPSS Statistics for Windows, Version 26.0, Armonk, NY, USA), via chi-square tests to examine differences between men and women in categorical variables including age, race/ethnicity, household income, education, and marital status. Additionally, chi-square tests were also utilized to examine sex differences between PTSD, substance use disorder, AUD, and DUD diagnoses, across absent, new, remitted, and ongoing diagnostic classifications.

For remaining analyses, data were analyzed using PROC SURVEYLOGISTIC in SAS, version 9.4 (SAS v9.4, SAS Institute Inc., Cary, NC, USA). All models were estimated using binomial distributions with logit link functions. Age, race, income, marital status, and education level were included as covariates. All models analyzed in SAS incorporated survey weights and stratification variables to account for the complex survey design.

The models regressed retrospective transitions in substance use disorders, AUD, and DUD, and sex onto retrospective transitions in PTSD. All participants were included in the analytic sample. Separate models were estimated for new vs. absent and ongoing vs. remitted PTSD diagnoses. The second set of models also regressed retrospective transitions in PTSD and sex onto retrospective transitions in SUD, AUD, and DUD. Similar to the above-mentioned analysis, all participants were included in the analytic sample and separate models estimated new vs. absent and ongoing vs. remitted SUD, AUD, and DUD diagnoses.

Additionally, based upon recent calls to study sex differences between men and women in regards to substance use (Becker et al., 2016, 2017; McHugh et al., 2018; Sanchis-Segura & Becker, 2016), it was decided a priori to create sex-specific models to explore the hypothesized associations within each sex group (as included in the primary hypothesis of this study). We chose to conduct these sex-specific models non-contingently on significant sex X PTSD/substance use disorder interactions based on the ample preexisting evidence for sex differences in the symptomatology of both PTSD and SUD. In light of recent reviews identifying sex differences in psychopathology as an understudied area contributing to health disparities (Becker et al., 2016, 2017; Sanchis-Segura & Becker, 2016), studies of such contingent sex differences are needed to identify sex differences in men and women with concurrent PTSD and SUD. As such, sex-specific models were also estimated, to characterize any sex differences between SUD/AUD/DUD on retrospective transitions in PTSD diagnosis and vice versa, as identified in the primary hypotheses of this study (Clayton & Tannenbaum, 2016; Ellis et al., 2020).

Results

Demographic characteristics by sex are reported in Table 1. All chi-square analyses found significant differences between males and females (p < 0.001), except age (p < 0.07). Additionally, chi-square analyses examining substance use diagnosis and PTSD diagnosis were significantly different between males and females (p < 0.001), with women experiencing more PTSD diagnoses in the past year than males and more males meeting criteria for the past year AUD, DUD, and substance use disorder diagnoses than females (see Table 2).

Table 1.

Demographics.

	Male	Female	χ 2	df	p-Value
Age (%)a			5.45	2	0.07
18–29 years old	22.9	21.9
30–44 years old	27.5	28.2
45+ years old	49.5	49.9
Race/Ethnicity (%)a			47.08	4	<0.001
White, non-Hispanic	53.9	52.0
Black, non-Hispanic	19.9	22.6
Native American, non-Hispanic	1.3	1.5
Asian, non-Hispanic	5.4	4.6
Hispanic, any race	19.5	19.3
Household Income (%)a			304.42	3	<0.001
$9999 or less	8.7	10.9
$10,000–$24,999	22.2	27.8
$25,000–$49,999	27.7	27.7
Over $50,000	41.4	33.6
Education (%)a			28.09	2	<0.001
Less than high school	15.5	14.8
Completed high school	28.2	26.1
Some college or higher	56.4	59.1
Marital status (%)a			693.40	5	<0.001
Married	42.2	37.9
Living with a partner	6.9	6.0
Widowed	3.7	9.8
Divorced	13.0	15.6
Separated	3.5	5.0
Never Married	30.5	25.7

df: degrees of freedom.

^aUnweighted values.

Table 2.

Sex differences in diagnoses.

	Male	Female	χ 2	df	p-Value
PTSD diagnosis (n)a			221.03	3	<0.001
Absent	15,184 (95.7)	18,786 (91.9)
New	40 (0.3)	80 (0.4)
Remit	154 (1.0)	406 (2.0)
Ongoing	484 (3.1)	1175 (5.7)
SUD diagnosis (n)a			898.20	3	<0.001
Absent	9798 (61.8)	15,581 (76.2)
New	1300 (8.2)	1002 (4.9)
Remit	2747 (17.3)	2375 (11.6)
Ongoing	2017 (12.7)	1489 (7.3)
AUD diagnosis (n)a			838.21	3	<0.001
Absent	10281 (64.8)	16027 (78.4)
New	1274 (8.0)	942 (4.6)
Remit	2623 (16.5)	2245 (11.0)
Ongoing	1684 (10.6)	1233 (6.0)
DUD diagnosis (n)a			263.64	3	<0.001
Absent	13,860 (87.4)	18,901 (92.4)
New	415 (2.6)	278 (1.4)
Remit	1114 (7.2)	917 (4.5)
Ongoing	443 (2.8)	351 (1.7)

df: degrees of freedom; SUD: substance use disorder; AUD: alcohol use disorder; DUD: drug use disorder.

^aUnweighted values.

New DSM-5 PTSD diagnoses and SUDs

New diagnoses of PTSD (vs. absent) were more likely among females across individuals with substance use disorder, AUD, and DUD (OR range 1.93–2.03). Additionally, being diagnosed with new, remitted, or ongoing SUD, AUD or DUD was associated with a new diagnosis of PTSD (OR range 2.53–8.1). See Table 3 for complete results.

Table 3.

Odds ratios for main effects and interactions between sex, substance use diagnosis, and new vs. absent PTSD diagnoses.

	Main effects
	New vs. Absent PTSD dx (REF Absent PTSD dx)
	Odds Ratio	95% CI	p-value
Substance use disorder
Female	2.03	1.20–3.44	0.01
New SUD dx	3.14	1.58–5.26	<0.01
Remit SUD dx	3.51	1.75–7.05	<0.01
Ongoing SUD dx	5.28	2.81–9.92	<0.01
Sex-specific analyses
Males
New SUD dx	2.11	0.64–6.99	0.22
Remit SUD dx	3.59	1.16–11.13	0.03
Ongoing SUD dx	4.98	1.88–13.22	<0.01
Females
New SUD dx	3.98	1.66–9.55	<0.01
Remit SUD dx	3.37	1.55–7.30	<0.01
Ongoing SUD dx	5.39	2.53–11.51	<0.01
Alcohol use disorder
Female	1.94	1.23–3.34	0.02
New AUD dx	2.53	1.17–5.47	0.02
Remit AUD dx	3.94	1.94–8.00	<0.01
Ongoing AUD dx	3.67	1.88–7.19	<0.01
Sex-specific analyses
Males
New AUD dx	2.08	0.72–5.97	0.17
Remit AUD dx	4.08	1.33–12.49	0.16
Ongoing AUD dx	2.11	0.75–5.96	0.05
Females
New AUD dx	2.86	1.00–8.17	0.05
Remit AUD dx	3.59	1.66–7.78	<0.01
Ongoing AUD dx	5.33	2.43–11.68	<0.01
Drug use disorder
Female	1.93	1.15–3.25	0.01
New DUD dx	8.11	3.51–18.74	<0.01
Remit DUD dx	3.78	1.85–7.61	<0.01
Ongoing DUD dx	6.24	3.17–12.29	<0.01
Sex-specific analyses
Males
New DUD dx	10.32	3.12–34.18	<0.01
Remit DUD dx	4.49	1.72–11.74	<0.01
Ongoing DUD dx	2.37	0.55–10.21	0.25
Females
New DUD dx	5.54	2.21–13.86	<0.01
Remit DUD dx	3.06	1.12–8.37	0.03
Ongoing DUD dx	10.96	5.10–23.58	<0.01

SUD: substance use disorder; AUD: alcohol use disorder; DUD: drug use disorder; Dx: diagnosis; CI: confidence Interval.

Sex-specific analysis: new DSM-5 PTSD diagnoses and SUDs

Sex-specific analyses found that men with new PTSD diagnosis had higher odds of remitted and ongoing substance use disorders (OR = 3.59, 4.98, respectively) and new and remitted DUD (OR = 10.32, 4.49, respectively). All women with new PTSD diagnoses had higher odds of new, remitted, or ongoing substance use diagnoses (SUD, AUD, and DUD). See Table 3 for complete results.

Ongoing DSM-5 PTSD diagnoses and SUDs

Regarding ongoing (vs. remitted) past year PTSD diagnosis, there were few significant main effects. Individuals diagnosed with ongoing PTSD had higher odds of also having an ongoing DUD (OR = 2.10). See Table 4 for complete results.

Table 4.

Odds ratios for main effects and interactions between sex, substance use diagnosis, and ongoing vs. remit PTSD diagnoses.

	Main effects
	Ongoing vs. Remit PTSD dx; REF Remit PTSD dx)
	Odds Ratio	95% CI	p-value
Substance use disorder
Female	0.89	0.65–1.20	0.43
New SUD dx	1.32	0.78–2.22	0.29
Remit SUD dx	1.08	0.83–1.39	0.57
Ongoing SUD dx	1.43	0.98–2.08	0.07
Sex-specific analyses
Males
New SUD dx	2.34	0.69–5.05	0.10
Remit SUD dx	1.01	0.59–1.74	0.96
Ongoing SUD dx	1.26	0.66–2.38	0.48
Females
New SUD dx	1.06	0.59–1.89	0.85
Remit SUD dx	1.11	0.84–1.46	0.47
Ongoing SUD dx	1.54	1.02–2.32	0.04
Alcohol use disorder
Female	0.86	0.64–1.17	0.34
New AUD dx	1.04	0.62–1.72	0.89
Remit AUD dx	0.97	0.76–1.26	0.84
Ongoing AUD dx	1.26	0.82–1.95	0.29
Sex-specific analyses
Males
New AUD dx	2.20	0.94–5.14	0.07
Remit AUD dx	0.91	0.55–1.52	0.71
Ongoing AUD dx	1.03	0.51–2.08	0.94
Females
New AUD dx	0.74	0.41–1.33	0.31
Remit AUD dx	1.01	0.76–1.34	0.07
Ongoing AUD dx	1.45	0.905–2.31	0.12
Drug use disorder
Female	0.89	0.66–1.20	0.42
New DUD dx	1.08	0.56–2.09	0.82
Remit DUD dx	1.12	0.85–1.49	0.41
Ongoing DUD dx	2.10	1.31–3.36	<0.01
Sex-specific analyses
Males
New DUD dx	0.63	0.11–1.18	0.35
Remit DUD dx	1.33	0.76–2.32	0.35
Ongoing DUD dx	1.75	0.77–3.78	0.18
Females
New DUD dx	1.72	0.24–2.12	0.24
Remit DUD dx	1.00	0.72–1.39	0.99
Ongoing DUD dx	2.40	1.28–4.53	0.007

SUD: substance use disorder; AUD: alcohol use disorder; DUD: drug use disorder; dx: diagnosis; CI: confidence interval.

Sex-specific analysis: ongoing DSM-5 PTSD diagnoses and SUD

Sex-specific analyses showed that women were almost 2.5 times more likely to have an ongoing DUD. No additional sex-specific analyses were significant. See Table 4 for complete results.

New DSM-5 SUD diagnoses and PTSD

New (vs. absent) diagnoses of a substance use disorder, AUD, and DUD were less likely among females (OR range 0.49–0.53) with PTSD. Additionally, being diagnosed with new, remitted, or ongoing PTSD was associated with a new diagnosis of SUD (OR = 3.16), AUD (OR = 2.50), and DUD (OR = 8.22). Remitted PTSD was associated with a new diagnosis of DUD (OR = 3.05). Ongoing PTSD was associated with an increased likelihood of a new SUD (OR = 1.88), AUD (OR = 1.68), and DUD (OR = 3.12) diagnosis. See Table 5 for complete results.

Table 5.

Odds ratios for main effects and interactions between sex, PTSD, and new vs. absent SUD, AUD and DUD diagnoses.

	Main effects (new vs. absent SUD dx; REF absent SUD dx)			Main effects (new vs. absent AUD dx; REF absent AUD dx)			Main effects (new vs. absent DUD dx; REF absent DUD dx)
	Odds Ratio	95% CI	p-Value	Odds Ratio	95% CI	p-Value	Odds Ratio	95% CI	p-Value
PTSD diagnosis
Female	0.53	0.47–0.59	<0.001	0.52	0.46–0.58	<0.001	0.49	0.40–0.61	<0.001
New PTSD dx	3.16	1.59–9.31	0.001	2.50	1.17–5.37	0.019	8.22	3.92–17.23	<0.001
Remit PTSD dx	1.41	0.88–2.28	0.15	1.61	1.03–2.53	0.039	3.05	1.65–5.65	<0.001
Ongoing PTSD dx	1.88	1.50–2.37	<0.001	1.68	1.33–2.12	<0.001	3.12	2.21–4.40	<0.001
Sex-specific analyses
Males
New PTSD dx	2.96	0.99–8.89	0.05	2.63	0.93–7.41	0.07	13.37	4.85–36.88	<0.001
Remit PTSD dx	0.62	0.23–1.66	0.34	0.69	0.33–1.47	0.34	3.77	1.56–9.12	0.004
Ongoing PTSD dx	1.90	1.27–2.86	0.002	1.90	1.27–2.84	0.002	2.80	1.64–4.78	<0.001
Females
New PTSD dx	3.27	1.35–7.90	0.009	2.42	0.84–7.02	0.10	4.86	1.93–12.21	<0.001
Remit PTSD dx	1.95	1.20–3.19	0.008	2.25	1.38–3.66	0.001	2.44	0.96–6.21	0.06
Ongoing PTSD dx	1.88	1.38–2.56	<0.001	1.56	1.14–2.14	0.007	3.30	2.03–5.95	<0.001

Sex-specific analysis: new DSM-5 SUD diagnoses and PTSD

Sex-specific analyses found that men with a new PTSD diagnosis had higher odds of new substance use disorder and DUD diagnoses (OR = 2.96, 13.37, respectively). Additionally, men with remitted PTSD were 3.77 times more likely to have a new DUD diagnosis. Men with ongoing PTSD were also more likely to like to have new SUD, AUD, and DUD (OR range 1.90–2.80).

Sex-specific analyses including women identified that those with new PTSD diagnosis had a higher likelihood of also being diagnosed with new substance use disorders (OR = 3.27) and DUD (OR = 4.86). Furthermore, those with remitted PTSD, as well as those with ongoing PTSD were more likely to have new SUD, AUD, and DUD (ORs range 1.95–3.30). See Table 5 for complete results.

Ongoing DSM-5 SUD, AUD, and DUD diagnoses and PTSD

Regarding ongoing (vs. remitted) past year substance use disorder, AUD, and DUD diagnoses, females were less likely to have ongoing SUD (OR = 0.80) and AUD diagnoses (OR = 0.80). Additionally, those with ongoing PTSD had a higher likelihood of also having ongoing SUD and DUD (ORs 1.40, 1.75, respectively).

Sex-specific analysis: ongoing DSM-5 SUD diagnoses and PTSD

The sex-specific analysis found that men with ongoing PTSD were more likely to have ongoing substance use disorders and DUD (OR = 1.49 and 1.56, respectively). Additionally, the female sex-specific analysis found that women with new PTSD diagnosis had higher odds of ongoing DUD (OR = 3.08) and those women with ongoing PTSD also had higher odds of ongoing SUD and DUD (OR = 1.34 and 1.93, respectively). No additional sex-specific analyses were significant. See Table 6 for complete results.

Table 6.

Odds ratios for main effects and interactions between sex, PTSD, and ongoing vs. remitted SUD, AUD, and DUD diagnoses.

	Main effects (ongoing vs. remitted SUD dx; REF Remit SUD dx)			Main effects (ongoing vs. remitted AUD dx; REF remit AUD dx)			Main effects (ongoing vs. remitted DUD dx; REF remit DUD dx)
	Odds Ratio	95% CI	p-Value	Odds Ratio	95% CI	p-Value	Odds Ratio	95% CI	p-Value
PTSD diagnosis
Female	0.80	0.70–0.91	0.001	0.80	0.70–0.92	0.003	0.96	0.77–1.19	0.68
New PTSD dx	1.39	0.72–2.66	0.33	0.87	0.98–2.03	0.75	1.34	0.60–2.98	0.47
Remit PTSD dx	1.06	0.78–1.49	0.15	0.86	0.59–1.27	0.45	0.93	0.60–1.45	0.75
Ongoing PTSD dx	1.40	1.15–1.69	0.001	1.14	0.93–1.40	0.20	1.75	1.31–2.33	<0.001
Sex-specific analyses
Males
New PTSD dx	1.52	0.50–4.66	0.46	0.57	0.12–2.74	0.48	0.37	0.08–1.70	0.20
Remit PTSD dx	1.09	0.60–1.97	0.78	0.86	0.45–1.66	0.66	1.04	0.51–2.12	0.92
Ongoing PTSD dx	1.49	1.13–1.95	0.005	1.11	0.82–1.49	0.50	1.56	1.06–2.31	0.03
Females
New PTSD dx	1.27	0.64–2.51	0.49	1.20	0.56–2.57	0.65	3.08	1.45–6.54	0.004
Remit PTSD dx	1.04	0.71–1.52	0.84	0.87	0.59–1.28	0.47	0.88	0.47–1.74	0.71
Ongoing PTSD dx	1.34	1.05–1.70	0.02	1.67	0.90–1.51	0.24	1.93	1.28–2.91	0.002

PTSD: posttraumatic stress disorder; SUD: substance use disorder; AUD: alcohol use disorder; DUD: drug use disorder; Dx: diagnosis; CI: confidence interval.

Discussion

The present study confirmed the bidirectional relationship between retrospective transitions of PTSD and substance use disorders in both men and women. Specifically, individuals with new, remitted, or ongoing SUDs, including AUD and DUD, experienced a higher likelihood of new PTSD diagnoses. Conversely, individuals with new, remitted, or ongoing PTSD showed a higher likelihood of new SUD diagnoses.

DSM-5 SUDs diagnoses and their relationship with new PTSD diagnoses

The present results demonstrated that individuals with new, ongoing, or remitted substance use disorders were 2.5–8 times more likely to develop a new diagnosis of PTSD. Additionally, both ongoing and remitted SUDs, AUDs, and DUDs were associated with new diagnoses of PTSD, indicating that preceding substance use may place individuals at risk for exposure to traumatic events. These findings provide further support that SUDs may place individuals at an increased vulnerability to trauma exposure and thus subsequent development of PTSD (Vujanovic & Back, 2019). It is important to note that the significant relationship between SUDS and new PTSD, remained even when the SUD was remitted. This illustrates that additional factors may moderate this relationship, which warrants additional research.

SUDs diagnoses and new PTSD diagnoses are moderated by sex

Sex-specific analysis revealed that women with a new PTSD diagnosis were more likely to have higher odds of all SUD diagnoses, including AUD and DUD, whereas men only had higher odds of remitted/ongoing substance use disorder and new/remitted DUD. This supports previous findings that substance use often mediates trauma symptoms and PTSD, especially among women (Berenz et al., 2019; Torchalla & Nosen, 2019). However, it is interesting that AUD among men was not associated with new PTSD. The absence association between AUD and new PTSD may highlight a women’s propensity to problematically consume alcohol to regulate negative affect and stress compared to their male counterparts (Peltier et al., 2019). Additional research should be directed to study this potentially important sex difference among men and women with new PTSD and alcohol use.

DSM-5 PTSD diagnoses and their relationship with new SUD diagnoses

PTSD diagnoses were associated with at least 2.5 times the likelihood of a substance use disorder, while individuals with ongoing PTSD also had a higher likelihood of an ongoing SUD or DUD. This provides evidence that the onset of new substance use may also occur to cope with PTSD symptomatology, as one may seek an “escape” from PTSD symptoms (Berenz et al., 2019; Jacobsen et al., 2001). This is further emphasized by the finding that individuals with ongoing PTSD had almost 1.5 times the likelihood of having an ongoing SUD or DUD.

PTSD diagnoses and new SUD diagnoses are moderated by sex

Interestingly, men with a new PTSD diagnosis were 3–15 times more likely to also have a new substance use disorder or DUD, and men with remitted PTSD were more likely to have a new DUD. Men with ongoing PTSD were more likely to have a new SUD and men with ongoing PTSD were more likely to report ongoing SUD and DUD. While previous data have illustrated that men are more likely to have a SUD precipitating PTSD (Back et al., 2005; Cottler et al., 2001), the present study illustrated a potentially temporal relationship between PTSD preceding or concurrently occurring with SUDs in men. Previous research has found that PTSD symptoms are often associated with a lack of awareness of emotions, as well as decreased control of behaviors, with men exhibiting markedly impaired awareness of emotions (Bornovalova et al., 2009), which may plausibly lead to problematic substance use to “numb” negative emotions. Thus, additional research examining the role of distress as a mechanism of underlying concurrent PTSD/SUD is needed.

Women with new PTSD were 3–4 times more likely to have a new substance use disorder or DUD and those women with ongoing PTSD were more likely to have a new SUD, AUD, or DUD. Furthermore, new and ongoing PTSD was associated with ongoing DUD among women. This is consistent with previous literature demonstrating that women are more likely than males to use substances for negative reinforcement (Becker & Koob, 2016; Peltier et al., 2019). While PTSD symptoms are highly associated with substance use in both men and women, women’s quantity of use is associated with coping (Lehavot et al., 2014), placing women at increased risk for problematic substance use to cope with negative affect. Women are also more likely to initiate substance use later in their lives, but progress to dependence quicker than their male counterparts (McHugh et al., 2018), thus potentially accounting for the association between PTSD and SUD diagnoses in this sample.

Sex differences in PTSD and SUD diagnoses

Women endorsed more PTSD diagnoses, spanning new, remitted, and ongoing diagnoses when compared to men. Conversely, a larger percentage of men met diagnostic criteria across new, remitted, and ongoing substance use disorders, including AUDs and DUDs. This finding supports previous literature that men meet diagnostic criteria for SUDs at higher rates than their female counterparts, while women endorse high rates of PTSD within the context of concurrent substance use (Grant et al., 2015, 2016). However, recent literature has demonstrated increasing rates of AUDs and DUDs as well as escalating rates of frequency and severity of substance use among women (Grant et al., 2015, 2016, 2017; Keyes et al., 2008; Seedat et al., 2009) Thus, ongoing research is needed to evaluate how such trends may impact the rates of SUDs among women with comorbid PTSD, especially as women are more likely to use substances for negative reinforcement of effect. Specifically, as substance use becomes more prevalent among women, likely, using substances to cope with effect will also increase.

Strengths/limitations

First, NESARC data is cross-sectional and thus we are unable to discern if a substance use disorder diagnoses preceded retrospective transitions in PTSD diagnosis or vice versa within the past 12 months. However, this research is the first, to the authors’ knowledge, to explore the sex differences in the role of concurrent PTSD and SUD and thus may inform future studies. Further, while the dataset is cross-sectional in nature, the categories were determined based upon changes noted by the participants. Participants indicated if there had been changes in diagnoses in the past year (e.g., within the past 12 months) or before the past year (e.g., <12 months ago), which allowed us to assess for current and prior diagnoses within the same interview. Also, NESARC-III queried participants regarding “Stressful Life Experiences” that they experienced or witnessed/learned about. This may lead to an under-diagnosis of PTSD, as it may not be a comprehensive evaluation of Criterion A events needed for the DSM diagnosis of PTSD. Additional cohort and longitudinal research with a comprehensive assessment of Criterion A events for PTSD diagnoses are warranted to continue to explore the pathways of concurrent PTSD and substance use to develop more targeted treatments. Finally, given the evidence that trauma-type may influence PTSD symptoms, future studies would benefit from also exploring the differences between Criterion A trauma types in PTSD/SUD populations.

The present analyses utilized sex-specific models to explore differences in new, remitted, and ongoing substance use disorder/AUD/DUD regardless of the presence of main effects or a significant interaction effect between sex and diagnosis. This was done to provide valuable data regarding the contingent sex differences present in SUDs and PTSD. Future studies would benefit from being powered to significantly explore such important sex differences. It also should be noted that, despite the relatively large sample, the population within this sample who met the criteria for PTSD was relatively small, and furthermore this sample was divided by sex and SUD comorbidity. Nonetheless, the present study resulted in large differences in point estimates between men and women, which is consistent with the hypothesized sex differences in these associations. The present study highlights important sex differences in retrospective transitions in new PTSD along with comorbid SUDs, but more research is needed with larger and more targeted samples to derive more precise estimates of sex differences.

Conclusions

Results indicate that new, remitted, and ongoing substance use disorders are related to new PTSD diagnoses, and new, remitted, and ongoing PTSD diagnoses are related to new SUDs, illustrating an important bidirectional relationship between PTSD and SUDs. Such associations were particularly strong among women, who have a heightened vulnerability to using substances to cope with negative affect and stress, which can be associated with both PTSD and ongoing substance use.

Data availability statement

The data that support the findings of this study are available with permission from NIAAA at https://www.niaaa.nih.gov/research/nesarc-iii.