Stress, sex, and symptomics: Individual posttraumatic stress symptoms and sexual behaviors in adolescents

Abstract

Posttraumatic stress disorder (PTSD) is linked with sexual risk, but less is known about which aspects of posttraumatic psychopathology may be most relevant in these associations. In trauma-exposed adolescents from a national sample (N=384, M_age =16.4, 72.4% female), we applied a novel “symptomics” approach—which focuses on the role of individual symptoms of psychopathology—to examine associations between PTSD symptoms and two sexual behaviors linked with potential risk: age at first sex and number of recent partners. In relative importance analyses, individual PTSD symptoms explained more variance in these sexual behaviors than did disorder diagnosis. For both sexual behaviors, sense of foreshortened future accounted for the greatest amount of variance; sleep-related symptoms (e.g., nightmares, sleep disturbance) also emerged as relevant. Additionally, dysphoria-related symptoms (e.g., concentration difficulties, anhedonia, restricted affect) explained significant proportions of age at first sex, whereas symptoms capturing trauma-related fear (e.g., intrusions, cued reactivity, hypervigilance) explained variance in number of partners. Findings suggest that select post-trauma symptoms are differentially associated with sexual behaviors in trauma-exposed adolescents. Screening for and intervening on these specific PTSD symptoms may mitigate possible sexual risk in this population, though additional research in larger and more sex-balanced samples is warranted.

Introduction

Trauma exposure and related psychological consequences—namely, posttraumatic stress disorder (PTSD)—are associated with poor sexual health and dysfunction across the lifespan (Yehuda et al., 2015), including during adolescence (Wang et al., 2019). Trauma and its psychiatric sequelae are not uncommon in youth; nearly two-thirds of adolescents in the U.S. will experience a traumatic event prior to age 17 (McLaughlin et al., 2013), and approximately 7% will have had diagnosable PTSD by the end of adolescence (Merikangas et al., 2010). Sexuality and sexual behavior also become increasingly important during this developmental period, and safe sexual behaviors—those that minimize risk of sexual violence victimization, unintended pregnancy, and the contraction of sexually transmitted infections (STIs)—can promote adolescents’ sexual health during this transitory time. Indeed, U.S. adolescents are disproportionately represented in STI prevalence rates (Centers for Disease Control and Prevention, 2021), and teen pregnancy is associated with adverse health outcomes for mothers (Sedgh et al., 2015). While youth sexual behavior in and of itself is not inherently problematic, addressing associations between trauma and sexual behaviors that may potentiate risk in adolescents could reveal important targets to improve sexual health and wellness—and thus foster both psychological and sexual well-being (Harden, 2014)—in this population.

While research has linked specific sexual (e.g., sexual abuse) and non-sexual (e.g., physical and emotional abuse, neglect) traumatic experiences to adverse sexual sequelae in adolescents (Miller et al., 2018; Thompson et al., 2017; Thomas et al., 2023), growing evidence supports posttraumatic psychopathology as a potential explanatory mechanism. Indeed, several papers have supported associations between trauma-related distress—conceptualized either categorically as a PTSD diagnosis or continuously as trauma-related symptom counts—and behaviors suggestive of potential sexual risk in youth. One study, conducted in the same epidemiological sample as in the present report, found that adolescents with PTSD reported a greater number of sexual partners compared to their non-diagnosed, yet trauma-exposed, counterparts (Miller et al., 2018). Similarly, one longitudinal study of youth at-risk for maltreatment found that trauma symptoms mediated associations between diverse forms of childhood trauma (e.g., sexual abuse, emotional abuse) and unprotected sex in adolescence (Thompson et al., 2017). However, PTSD is heterogenous (Zoellner et al., 2014), in that there are over 636,000 ways to meet diagnostic criteria (Galatzer-Levy & Bryant, 2013), and even subthreshold manifestations of the disorder—that is, clinical presentations that do not meet full diagnostic criteria—are associated with severe distress and impairment (McLaughlin et al., 2015). While informative at a broad level, studying trauma-related psychopathology conceptualized as a diagnosis or total count score limits understanding of which components of posttraumatic stress may be most relevant to sexual behavior.

Approaching posttraumatic psychopathology dimensionally may shed light on more granular relationships. Indeed, PTSD can be conceptualized along dimensions of fear and dysphoria (Zoellner et al., 2014), with discrete subdimensions of symptoms nested beneath these broader constructs. A wide body of factor analytic research supports the decomposition of PTSD into underlying symptom dimensions (see Armour et al., 2016 for a review), with recent research supporting as many as seven distinct dimensions of similar symptoms: re-experiencing, avoidance, negative affect, anhedonia, externalizing behaviors, anxious arousal, and dysphoric arousal (Armour et al., 2015). Symptom dimensions that reflect threat reactivity or elicit a fear response—such as re-experiencing, avoidance, and hyperarousal symptoms—can be considered under the higher-order fear dimension, while dimensions that are more dysphoric in nature—such as negative affective or dysphoric arousal—are captured by dysphoria. Research supports differential associations between both these lower-order and higher-order symptom dimensions and various health outcomes (El-Gabalawy et al., 2018; Sumner et al., 2020).

Within this area of research, a relatively small literature has considered PTSD symptom dimensions and sexual health and functioning, with most research studying these associations in circumscribed samples of reproductive-aged women (e.g., military populations, as reviewed by Bird et al., 2021). Less is known about links between PTSD symptoms and sexual behavior in youth, who are overrepresented in adverse sexual sequelae and for whom certain types of sexual behaviors may be particularly impactful. To date, only two studies have examined these questions in adolescents and young adults, findings from which converged on the importance of trauma-related intrusions (e.g., re-experiencing symptoms) for risky sexual behaviors such as unprotected or substance-involved intercourse (Banks et al., 2022; Walsh et al., 2014). Notably, both of these samples were comprised of substance-dependent youth, who may differ from trauma-exposed adolescents who do not use substances.

Given differential associations observed across symptom dimensions, it follows that select posttraumatic stress symptoms—that is, the individual PTSD symptoms that comprise these larger dimensions—may be particularly linked with certain sexual behaviors that could increase risk. Increasing research in psychopathology—primarily examining symptoms of depression and PTSD—has adopted a “symptomics” approach, one that focuses on the associations between select psychopathology symptoms and various functional outcomes (Fried & Nesse, 2014; Kachadourian et al., 2019; Kachadourian et al., 2021). This literature has illustrated that certain symptoms of psychopathology contribute more strongly to select aspects of functioning than others. For example, individual PTSD symptoms were differentially linked with physical and mental health in veterans; while concentration difficulties contributed strongly to mental dysfunction, this symptom was less relevant to physical functioning, for which anhedonia and sleep were more impactful (Kachadourian et al., 2019). Another symptomics study of adults found sleep and concentration difficulties were of particular relevance to work and home activities, though of less relevance to relationships than anhedonia or low mood (Freid & Nesse, 2014). Collectively, this work highlights the relevance of dysphoria-related symptoms across domains.

This symptomics methodology is also well-aligned with theoretical frameworks that de-emphasize diagnostic models in psychiatry in favor of trans- or non-diagnostic approaches that more strongly focus on functioning. For example, the Power Threat Meaning Framework (Johnstone & Boyle, 2018) conceptualizes trauma-related symptomatology—that is, an individual’s behaviors and emotional reactions following trauma—as understandable ways in which to cope with threat. Toward this way of thinking, trauma-related symptoms are highly contextual, frequently functioning as adaptative responses to stress and adversity, and can be considered as reasonable versus solely psychopathological reactions warranting diagnosis. At the same time, the model recognizes that the responses that promoted survival in certain contexts can become less functional in others—thus signaling the value of an approach that considers the functional implications of these trauma-related symptoms from a non-diagnostic lens. Separating symptoms from disorders is also a hallmark feature of the National Institute of Mental Health’s Research Domain Criteria (RDoC), a research framework that focuses on core functional domains (e.g., negative valence, cognitive systems, social processes) that cut across diagnoses (Insel et al., 2010). Studying symptom-level associations with sexual risk specifically may illuminate those symptoms that are particularly relevant to functioning, including those potentially driving the aforementioned dimension-specific findings. Such results could inform identification of focused intervention targets for trauma-exposed youth more broadly.

In this study, we applied a novel “symptomics” approach to examine which PTSD symptoms accounted for the greatest proportion of variance—while accounting for intercorrelations between symptoms—in two sexual behaviors (age at first sex, number of recent sexual partners) in trauma-exposed youth. While these sexual behaviors are not necessarily risky in and of themselves, both younger age at first sex and a greater number of partners have been linked with sexual risk factors such as inconsistent contraceptive use (Ashenhurst et al., 2016; Heywood et al., 2015)—which in turn can increase opportunities for STI contraction or unintended pregnancy. Better understanding of the behaviors that have been associated with sexually risky actions may illuminate potential intervention targets; thus, we focus on the sexual behaviors of age at first sex and number of recent partners. We hypothesized that dysphoria-related symptoms—including sleep and concentration problems—would be most relevant to these behaviors in this national, non-clinical sample of adolescents.

Method

Participants and Procedure

Data were from the National Comorbidity Survey Adolescent Supplement (NCS-A), a national cross-sectional survey of 10,184 U.S. adolescents (Kessler et al., 2009). From 2001–2004, youth were recruited from households and schools. Adolescents recruited from households (n=879) were sampled from homes that participated in the National Comorbidity Survey Replication (NCS-R), a national survey of adult psychopathology (Kessler & Merikangas, 2004); the remaining participants were recruited from a representative sample of schools in the NCS-R’s data collection area (n=9,244). Cases were weighted to approximate the 2000 Census.

Youth completed face-to-face interviews with trained research staff, while their guardians completed questionnaires. Measures queried youths’ medical and mental health and risk and protective factors for psychopathology. Written assent and consent were obtained from adolescents and their caregivers, respectively, and participants were paid $50. All NCS-A procedures were approved by the Human Subjects Committees of the University of Michigan and Harvard Medical School. More information on the NCS-A study design and methods is available elsewhere (Kessler et al., 2009).

A total of 1,064 youth in the NCS-A (10.5%) reported on PTSD symptoms, and 944 had complete data for all symptoms. Given our interest in the role of individual PTSD symptoms, we restricted analyses to the 944 participants who had complete data across symptoms. Of these youth, 384 reported a history of sexual activity, as ascertained by affirmative responses to the following question: “Have you ever had sexual intercourse?” These individuals had available data on age at first sex and number of past-year partners and thus constituted the final analytic sample.

Measures

Exposure to potentially traumatic events and PTSD.

Adolescents completed a modified Composite International Diagnostic Interview (CIDI; Kessler & Üstun, 2004). Via this diagnostic interview, adolescents indicated if they ever experienced one of 19 potentially traumatic events consistent with a Criteria A1 exposure (e.g., interpersonal violence, accident, witnessing event). Youth were also asked to indicate if they had experienced a traumatic event other than those specifically prompted, as well as invited to select another “private” event if they did not wish to disclose details about the experience. Those who reported one or more events were then asked if these experiences were associated with posttraumatic symptoms such as upsetting memories or dreams, feeling emotionally distant or depressed, experiencing difficulties sleeping or concentration, and feeling jumpy or startling easily. Youth who reported more than one potential trauma exposure were asked to identify the event associated with the greatest number of these types of symptoms; this event was then utilized as the index trauma upon which PTSD symptoms were queried.

Aligned with prior work in NCS-A (McLaughlin et al., 2013), we collapsed these potentially traumatic events into interpersonal, non-interpersonal, and other experiences (see Table 1 for prevalence rates of these exposures). Interpersonal events included experiences of sexual violence, physical abuse (by partner, caregiver, or another individual), stalked, kidnapped, and witnessing domestic violence. Non-interpersonal events included the unexpected death of a loved one, serious or life-threatening accident such as a motor vehicle accident or a natural disaster, witnessing death or serious injury of another person, and learning of a traumatic event happening to a loved one. Other events encompassed those not previously described, as well as experiences youth preferred to keep private that were followed by posttraumatic stress symptomatology. More information on trauma exposure in NCS-A is published elsewhere (McLaughlin et al., 2013).

Table 1.

Index Traumas Endorsed by the Analytic Sample (N=384)

	% (N)
Interpersonal	38.3 (147)
Sexual violence	20.6 (79)
Physical abuse	9.1 (35)
Stalked	2.6 (10)
Kidnapped	0.3 (1)
Witnessed domestic violence	5.7 (22)
Non-interpersonal	49.5 (190)
Unexpected death of loved one	33.3 (128)
Serious or life-threatening accident	9.1 (35)
Witnessing death or serious injury	5.5 (21)
Learned of traumatic event to loved one	1.6 (6)
Other	12.2 (47)
Some other event	6.3 (24)
Private event	6.0 (23)

PTSD.

PTSD was assessed via a modified Composite International Diagnostic Interview (CIDI; Kessler & Üstun, 2004). Adolescents indicated if they ever experienced a Criteria A1 trauma (e.g., abuse, disaster), with PTSD symptoms anchored to their “worst” experience. Because the NCS-A assessed PTSD using DSM-IV criteria, Criteria A2 was also collected (e.g., experienced fear, helplessness, shock, horror during the traumatic event). Given that this indicator does not significantly contribute to PTSD prevalence (Weathers & Keane, 2007), we excluded it from analyses. We studied the remaining 17 symptoms from Clusters B (re-experiencing), C (avoidance-numbing), and D (arousal), treating these binary indicators of lifetime symptom history as individual predictors in symptomics analyses. Consistent with work in the NCS-R (e.g., Sareen et al., 2005), we also summed the number of endorsed symptoms to create a total PTSD symptom count. Adolescents were also categorized by lifetime PTSD diagnosis (yes/no) if they met DSM-IV diagnostic criteria.

Responses indicated lifetime symptom presence (yes/no) across the 17 DSM-IV symptoms of PTSD. In addition to utilizing these indicators as individual predictors in symptomics analyses, we calculated a total PTSD symptom count and categorized youth by diagnosis (yes/no).

Sexual behaviors.

Adolescents who endorsed sexual experience at the time of the interview reported on their age at first sexual intercourse (n=364). Consistent with prior work in NCS-A (Thomas et al., 2023; Vafai et al., 2020), we excluded youth who reported age at first sex ≤11 years in an effort to capture consensual activity only. Youth with sexual experience also indicated their number of past-year sexual partners (n=380). Because of extreme outliers on this indicator (e.g., 90 partners), we winsorized the top 1% of the distribution. Additional data considering the parameters of these sexual encounters (e.g., coerced versus consensual, protected versus unprotected) were not available.

Sociodemographics.

Participant sex (female, male) and age in years were included as covariates, following previous symptomics research (Fried & Nesse, 2014; Kachadourian et al., 2019).

Data Analysis

Using survey-weighted regression (Lumley, 2004), we examined associations between PTSD diagnosis, total symptom count, and the sexual behaviors. We then applied a two-step “symptomics” approach to investigate whether adopting a symptom-level analysis afforded unique insights beyond a diagnosis or total symptom count. First, we considered all 17 PTSD symptoms as predictors and estimated R² values to determine the total amount of variance in each sexual behavior collectively accounted for by these symptoms. Second, we completed relative importance (RI) analyses (Groemping, 2007), which calculate the percentage that each predictor—here, the PTSD symptoms—contribute to the overall association with each outcome—here, each sexual behavior—while accounting for intercorrelations. The total R² is then divided into values of explained variance accounted for by each individual symptom; we scaled values to sum to 100% to aid interpretability. To assess robustness of results, we also conducted two sensitivity RI analyses, covarying 1) PTSD diagnosis to determine if associations were better explained by threshold PTSD and 2) sexual assault to investigate if this specific trauma type particularly accounted for associations given relevance for sexual risk.

Using the relaimpo package in R, we completed relative importance (RI) analyses (Groemping, 2006) to determine the percentage of variance accounted for in the sexual behaviors by each of the 17 PTSD symptoms while simultaneously accounting for intercorrelation. Using the LMG metric, the total R² is divided into values of explained variance accounted for by each individual symptom, resulting in one RI value per symptom.

Due to a robust literature linking sexual violence with later-life sexual risk (Noll, 2021), we considered sexual assault as a covariate in sensitivity analyses. Specifically, we replicated RI analyses that decomposed the variance in the sexual behaviors among the 17 PTSD symptoms while also accounting for if youths’ symptoms were anchored to an experience of sexual violence.

Results

Adolescents were about 16 years of age, on average (SD=1.14, range=13–18). Most were female (72.4%) and White (52.1%). Nearly one-half (48.2%) had a PTSD diagnosis, with an average of 10.4 symptoms (range=2–17; Table 2). The most common symptom was avoidance of trauma-related thoughts. Mean age at first sex was 14.7 years (SD= 1.46, range=12–18); the average number of past-year partners was 1.7 (SD=1.27, range=0–5). The most commonly occurring traumatic event was unexpected death of a close other (33.3%); sexual violence was the second most commonly endorsed experience, with approximately 20% of the sample reporting this event type as the index trauma (see Table 1).

Table 2.

Associations between posttraumatic stress and sexual behaviors (N=384)

	Mean (SD) or % (n)	Age at first sex	Number of past-year partners
PTSD Diagnosis Model		R2 = 0.26	R2 = 0.02
		β	β
Lifetime PTSD diagnosis	48.2 (185)	−0.36t	0.02
Sex (female)		0.56*	−0.42t
Age		0.73*	0.01
Total PTSD Symptom Count Model		R2 = 0.26	R2 = 0.02
		β	β
Total PTSD symptom count	10.43 (3.55)	−0.15	0.05
Sex (female)		0.55*	−0.44t
Age		0.72*	0.01
Individual PTSD Symptoms Model		R2 = 0.36	R2 = 0.15
		β	β
Intrusive thoughts	80.7 (310)	−0.16	−0.37
Nightmares	64.6 (248)	−0.11	0.45t
Flashbacks	65.9 (253)	0.11	−0.20
Cued psychological reactivity	82.0 (315)	0.06	0.20
Cued physiological reactivity	47.9 (184)	0.24	−0.31
Avoiding thoughts of trauma	91.4 (351)	0.31	0.04
Avoidance of trauma reminders	68.2 (262)	−0.23	0.21
Amnesia	33.3 (128)	−0.40*	−0.26
Loss of interest	44.0 (169)	−0.29	−0.21
Interpersonal detachment	56.5 (217)	0.12	0.14
Restricted affect	54.4 (209)	−0.39	0.10
Sense of foreshortened future	28.4 (109)	−0.19	0.54*
Sleeping problems	74.0 (284)	0.34t	−0.22
Irritability	55.7 (214)	−0.19	0.13
Difficulty concentrating	71.9 (276)	0.37	−0.54*
Hypervigilance	70.1 (269)	−0.25	0.20
Exaggerated startle	53.6 (206)	0.11	0.06
Sex (female)		0.35	−0.50*
Age		0.72*	0.02

Note: All models adjusted for participant sex (female, male) and age.

^*p≤.05,

^tp≤.10

As seen in Table 2, neither PTSD diagnosis nor total symptom count was significantly linked with age at first sex or number of partners. Models accounted for approximately one-quarter of the variance in age at first sex (R² = 0.26) and minimal variance in number of partners (R² = 0.02). Analyses including all 17 symptoms revealed that only three symptoms—amnesia (β=−0.40), sense of foreshortened future (β=0.54), and concentration difficulties (β=−0.54)—reached statistical significance across the indicators; others were at trend-level. The total amount of variance explained ranged from 15% (number of partners) to 36% (age at first sex).

In RI analyses, nearly 25% of the variance in age at first sex was accounted for by five symptoms: sense of foreshortened future (5.0%), sleeping problems (4.9%), irritability (3.8%), loss of interest (3.6%), and restricted affect (3.4%; Figure 1). Excluding sleeping problems, all were associated with younger age at first sex. In contrast, three symptoms—sense of foreshortened future (10.2%), nightmares (9.5%), and concentration difficulties (8.5%)—accounted for over 25% of the variance in number of partners (Figure 2). An additional 15% of the variance in this indicator was captured by intrusive thoughts (5.8%), psychological reactivity (5.5%), and hypervigilance (3.9%). Youth with concentration problems and intrusive thoughts reported fewer partners, while those with a sense of foreshortened future, nightmares, psychological reactivity, and hypervigilance reported more partners. Across models, PTSD diagnosis accounted for small (1.1–1.5%) proportions of variance. Sexual trauma also contributed minimally, though was more relevant to partner history (1.7%) than age at first sex (0.4%).

Figure 1. Relative importance coefficients of PTSD symptoms on age at first sex.

Values correspond to relative importance coefficients, or the unique shared variance between PTSD symptoms and age at first sex, controlling for sociodemographic covariates. Error bars represent 95% confidence intervals. Relative importance explained by age and sex was 5.25% and 64.60%, respectively. Covariates are not represented in figures for ease of interpretability.

Figure 2. Relative importance coefficients of PTSD symptoms on number of partners.

Values correspond to relative importance coefficients, or the unique shared variance between PTSD symptoms and number of past-year partners, controlling for sociodemographic covariates. Error bars represent 95% confidence intervals. Relative importance explained by age and sex was 1.31% and 40.20%, respectively. Covariates are not represented in figures for ease of interpretability.

Sensitivity analyses considering sexual trauma revealed that this particular experience did not significantly contribute to variance in sexual behavior, accounting for approximately half a percentage point to age at first sex (0.4%) and slightly more (1.7%) to number of past-year partners. Moreover, patterns of results did not significantly change when including this covariate; individual symptoms remained differentially associated with sexual behavior and at similar magnitudes across the indicators.

Discussion

PTSD is linked with sexual risk in youth, but less is known about which symptoms may be most relevant. Here, when examining PTSD as a diagnosis and total symptom count, associations with two behavioral indicators of potential sexual risk were mixed, with minimal to moderate amounts of overall variance accounted for by posttraumatic psychopathology. A more complete picture emerged when we adopted a “symptomics” approach, in that we observed differential associations between individual symptoms and sexual behaviors. Across both indicators, sense of foreshortened future accounted for the greatest proportion (5–10%) of variance; sleep-related symptoms such as general sleep disturbance and nightmares were also relatively more prominent than other symptoms, collectively accounting for 6–12% of variance in the sexual behaviors. Symptoms representing higher-order dysphoria—including but not limited to interpersonal detachment, restricted affect, irritability, limited concentration, and anhedonia—accounted for greater proportions of variance in age at first sex than did individual fear-related symptoms. Indeed, many fear-based items related to re-experiencing or avoidance of trauma-related stimuli did not account for substantial variance in this behavior, in some cases contributing less than one percent. In contrast, fear-related symptoms (e.g., intrusive thoughts, psychological reactivity, hypervigilance) were more relevant for number of partners, collectively accounting for over 25% of variance. Across all analyses, individual trauma-related symptoms accounted for a greater proportion of variance in sexual behavior than did threshold PTSD—even in a sample where nearly one-half of individuals met diagnostic criteria—thereby highlighting the value of a symptom-level approach. While some symptoms emerged as more relevant than others across the indicators, the proportion of variance accounted for in outcomes was relatively small, and no single symptom captured more than 10% of the variance in a particular behavior.

Our results are somewhat consistent with prior PTSD symptomics research examining other health indicators (Cleveland et al., 2022; Kachadourian et al., 2019), whereby symptoms of dysphoria proved particularly relevant. Indeed, in this study, the dysphoria indicator of foreshortened future—or expectation of a short life—emerged as especially pertinent. It may be that youth with this symptom engage in certain sexual behaviors in part because they do not anticipate living long enough to confront potential consequences. There may also be sociocultural or evolutionary pressures to reproduce, and the possibility of a foreshortened life could result in more frequent sexual activity to fulfill those social roles and expectations. These proposals align with Life History Theory’s proposal that early trauma influences reproductive development (Belsky, 2019). Alternatively, youth endorsing future-oriented negativity may be more likely to engage in risky behaviors—which sexual behaviors can be if not approached safely—in general, likely forming a feedback loop between risk-taking and existential dread.

Sleep-related symptoms—including both trauma-specific nightmares and general sleep difficulties—were also highly relevant, accounting for the second largest proportion of variance for both indicators. Importantly, the relevance of certain sleep-related experiences differed. While nightmares accounted for a large proportion of variance in number of partners, this symptom was less relevant for age at first sex, where general sleeping difficulties were more pertinent. The direction of these associations also differed, in that general sleeping problems were linked with older age at first sex, while youth who reported nightmares endorsed more partners. It may be that youth pursue multiple partnerships in efforts to suppress trauma-related nightmares, some of which may be anchored to prior sexual encounters (i.e., if a sexual trauma survivor). Alternatively, experiencing trauma-related nightmares could emotionally dysregulate youth, prompting them to pursue sexual relationships with others in service of increasing connection and engaging in co-regulation. In contrast, youth experiencing more generalized sleeping difficulties (e.g., staying asleep) may be less likely to pursue sexual relationships and thus initiate sex later in adolescence. It could also be that these youth are more engaged in other life domains of importance in adolescence (e.g., schoolwork, extracurriculars, friendships), participation in which can limit the amount of time available both for sleep and pursuit of sexual relationships.

Fear-related symptoms played a role in one, but not both, sexual behaviors. The associations observed between these specific symptoms and number of partners are somewhat consistent with prior work, though with some differences. In prior research sampling substance-using adolescents, the re-experiencing symptom dimension was linked with a greater number of sexual partners (Banks et al., 2022). We found that although many of the symptoms comprising this dimension were relevant to this indicator, the direction of effects varied: some re-experiencing symptoms, such as nightmares, were linked with a greater number of partners, while others—including intrusive thoughts—were associated with fewer partners. These differences were only observable when decomposing the re-experiencing dimension into its component parts, a more granular approach that may illuminate if specific symptoms drive links between PTSD and potentially risky sexual behaviors. Other work in young adults found that re-experiencing symptoms are associated with other sexual risk indicators, including sex with uncommitted partners (Walsh et al., 2014). However, here we did not observe these symptoms to be as impactful to age at first sex, for which dysphoria appeared more relevant. This patterning suggests that certain components of PTSD may be more relevant to discrete sexual behaviors. Discrepancies may also be due, in part, to methodological differences, including diverse operationalizations and samples, as prior work has exclusively studied substance-dependent youth and assessed sexual behaviors more concretely suggestive of risk (e.g., unprotected sex) than what we were able to examine here. Additionally, one prior study only measured intrusions (Walsh et al., 2014), thus limiting investigation of other symptoms’ contributions.

Similar to prior work on these topics (e.g., Banks et al., 2022; Walsh et al., 2014), our analytic sample was predominantly female. Given that females and males were equally represented in the epidemiologic survey from which these data are drawn, this imbalance is likely because we restricted the sample to only include individuals with complete data on the constructs of interest (i.e., all PTSD symptoms, sexual behaviors). The sex composition of our sample is an important consideration, given that sex differences in PTSD symptomatology often emerge during adolescence, with females exhibiting higher overall symptom levels than males (Haag et al., 2020) and observed inequities in PTSD symptom centrality or interconnectedness among youth (Cao et al., 2019). Indeed, a prior study considering symptom centrality identified foreshortened future as a highly interconnected PTSD symptom for male, but not female, youth (Cao et al., 2019). Though this particular symptom emerged as relevant to both sexual behaviors in our study, it accounted for significantly less variance (e.g., 5–10%) than did participant sex (40–65%) across the indicators. Relatedly, evidence suggests that traumatic stress amplifies sex differences in adolescent sexual behavior. While male youth typically endorse higher baseline levels of sexual activity, the association between PTSD symptoms and certain sexual behaviors is stronger among females (Cavanaugh, 2013). Future research could consider adopting a sex-sensitive perspective to investigate these questions in larger and more sex-balanced samples and to increase generalizability of findings.

Our study has limitations. Data were cross-sectional and assessed lifetime PTSD symptoms; thus, directions of effects cannot be determined. Longitudinal or ecological monetary assessment research is needed to establish temporality and clarify potential mechanisms. Given the research question, we only included trauma-exposed youth with complete PTSD symptom data and who reported sexual activity. Due to NCS-A’s timing, DSM-IV PTSD symptoms were assessed; this is an important consideration, as DSM-5 PTSD introduced new symptoms focused on negative alterations in cognitions and mood (e.g., blame, negative emotional state) that we were unable to analyze here. However, to our knowledge, NCS-A is the only national study to date with data on both PTSD symptoms and sexual behaviors. Future research could assess both a wider range of sexual behaviors, including those that are more explicitly sexually risky and those that may be more common in contemporary society (e.g., online behaviors), and current diagnostic criteria for PTSD. Moreover, we considered age at first sex and number of sexual partners as sexual behaviors that could indicate possible sexual risk, given that these behaviors, when not carefully managed, have the potential to precipitate unintended pregnancy and sexually transmitted infections among youth. However, as is typical with epidemiological data, we acknowledge that we do not know the context of these behaviors (e.g., unprotected versus protected sex), which may result in overestimations of risk and relevance. Relatedly, the total amount of variance accounted for in models was small, with small percentage contributions per symptom (e.g., 3–5%). Further, many predictors did not reach significance; however, predictors with non-significant coefficients can still contribute to explained variance due to intercorrelation, as is common with psychopathology symptoms. At the same time, the clinical significance of these associations must be interpreted in the context of small amounts of explained variance.

Limitations aside, this study is also characterized by strengths, including drawing data from a large, population-representative sample of adolescents, among whom adverse sexual health outcomes are overrepresented and can be particularly deleterious to health and development; PTSD symptom assessment via diagnostic interview, unlike most prior symptomics research, which has historically utilized self-report measures; and sensitivity analyses considering threshold PTSD and sexual trauma. Moreover, the novel symptomics approach—wherein the focus is on the functional impact of select psychiatric symptoms versus on psychiatric diagnoses—is synergistic with contemporary efforts in psychology to conceptualize psychopathological reactions on a spectrum. Indeed, RDoC and its dimensional view of psychopathology has been increasingly advocated for as an alternative to the categorical DSM, with growing research employing its use (Pacheco et al., 2022). Future research could continue to integrate these approaches and examine the impact of trauma-related symptoms on a variety of functional outcomes in non-clinical populations. Such research may shed light on underlying mechanisms linking trauma exposure with adverse sequelae across domains.

In sum, our study findings suggest that a PTSD diagnosis or total symptom count may be less informative than considering component aspects of posttraumatic psychopathology—whether through addressing PTSD symptom dimensions or individual items. Select PTSD symptoms—particularly sense of foreshortened future and sleep-related disturbances—may function to help drive potentially risky sexual behaviors in trauma-exposed youth. Future research could consider evaluating whether targeted intervention on these symptoms mitigates sexual risk during this important developmental window.

Availability of data and material:

The NCS-A data are publicly available at https://www.icpsr.umich.edu/web/HMCA/studies/28581. Copies of the code are available upon reasonable request to the first-author.