PSYCHOSOCIAL AND BEHAVIORAL CORRELATES OF ANXIETY SYMPTOMS IN A SAMPLE OF HIV-POSITIVE, METHAMPHETAMINE-USING MEN WHO HAVE SEX WITH MEN

Abstract

Studies show high rates of psychiatric symptoms among methamphetamine users; however, little information exists regarding methamphetamine use and anxiety. This study investigated psychosocial and behavioral correlates of anxiety symptoms in a sample of 245 HIV-positive men having sex with men (MSM) who were enrolled in a sexual risk reduction intervention. In a multiple regression analysis, anxiety symptoms were associated with homelessness, recent experience of HIV symptoms, injection drug use, lifetime sexual abuse, engaging in risky sexual behaviors, and seeking out partners at risky sexual venues when “high” on methamphetamine. These findings can be used to inform and refine sexual risk reduction interventions and substance use treatment programs for HIV-positive methamphetamine-using MSM.

Introduction

Anxiety symptoms and anxiety disorders are common in the general population and have been associated with disability, reduced quality of life, and a range of medical conditions, including diabetes, heart disease, and cancer (Gadalla, 2008; Mogotsi, Kaminer, & Stein, 2000; Muller, Koen, & Stein, 2005). Symptoms of anxiety include emotional (e.g., fear, apprehension, nervousness), cognitive (e.g., catastrophic thinking, worry, obsessive thoughts), and physiological (e.g., muscle tension, shallow breathing, heart racing, sweating) experiences (Derogatis & Melisaratos, 1983). The development of anxiety disorders is associated with genetic factors, brain chemistry, environmental factors, and psychological traits (Barlow, 2002). The severity of symptoms can influence patient response to both medications and psychological treatments.

Several studies have reported that anxiety symptoms are common among methamphetamine users (Nakama et al., 2008; Zweben et al., 2004); in particular, the acute phase of methamphetamine withdrawal is characterized by high levels of anxiety (McGregor et al., 2005). It has also been reported that methamphetamine users who have high levels of anxiety have more sexual risk behaviors and greater substance use morbidity compared to those with lower levels of anxiety (Lubman, Allen, Rogers, Cementon, & Bonomo, 2007). An important factor that influences anxiety symptoms appears to be HIV serostatus. Several studies report elevated levels of anxiety symptoms among HIV-infected patients (e.g., Peng et al., 2010; Kagee & Martin, 2010). It has been suggested that some HIV-positive individuals may use methamphetamine to reduce anxiety related to their HIV symptoms (Tsao, Dobalian, Moreau, & Dobalian, 2004).

To date, limited research has been done on the psychosocial and behavioral correlates of anxiety among methamphetamine users. Correlates of anxiety in this group may differ from those of other groups for at least two reasons. First, methamphetamine users may have greater exposure to stressful or anxiety-provoking environmental stimuli because of their drug-using lifestyle (e.g., criminal activity). Second, methamphetamine users may respond differently to stressful stimuli because of the drug's effects on brain chemistry and neurotransmitters. We would therefore expect to see a different pattern of correlates of anxiety symptoms in this group.

The psychiatric literature identifies several psychosocial and behavioral factors that have strong associations with anxiety disorders. Homelessness and marginal housing status have been associated with anxiety in homeless adolescents and injection drug users (IDUs) (Aichhorn et al., 2008; Waldrop-Valverde & Valverde, 2005). Studies in the general population have also reported an association between economic hardship, unemployment, and anxiety (Molarius et al., 2009; Scutella & Wooden, 2008).

High levels of anxiety have also been associated with injection drug use. In a study of anxiety in Hispanic IDUs, Reyes et al. (2007) reported that 37.1% had severe symptoms, and those with severe anxiety were more likely to engage in risky drug use behaviors. In another study of IDUs, anxiety was significantly associated with sharing needles (Lundgren, Amodeo, & Chassler, 2005). Accordingly, we posited that injection of methamphetamine would be associated with more anxiety symptoms than would other routes of administration.

There is also evidence that family functioning is associated with anxiety disorders (Bogels & Brechman-Toussaint, 2006). Negative qualities of the parent-child relationship, particularly conflict during adolescence, have been linked to anxiety and depression (Adams & Laursen, 2007). In our research, family conflict was common among adult methamphetamine users (Semple, Strathdee, Zians, & Patterson, 2009), suggesting the need to examine family conflict and anxiety in this population. Moreover, given that an inverse relationship has been found among HIV-positive individuals between social support and depression (Abramowitz et al., 2009; Mavandadi, Zanjani, Ten Have, & Oslin, 2009), we hypothesized that a similar relationship would be found between social support and anxiety.

Several studies have revealed that childhood sexual abuse (CSA) is associated with increased risk of anxiety symptoms and substance dependence in adulthood (Fergusson, Boden, & Horwood, 2008; Sun et al., 2008). It has also been documented that sexual assault on an adult male has implications for long-term psychological well-being, including elevated levels of anxiety symptoms (Anderson, 1981–2). Studies of the relationship between sexual risk behavior and anxiety symptoms have produced mixed findings. In a study of methamphetamine-dependent gay and bisexual men, Shoptaw, Peck, Reback, and Rotheram-Fuller (2003) reported that 54% of the sample met criteria for lifetime anxiety disorder, and those with anxiety disorder were found to have higher rates of genital gonorrhea, suggesting a positive relationship between anxiety and sexual risk. In contrast, Parsons, Halkitis, Wolitski, and Gomez (2003) reported that HIV-positive MSM who had unprotected receptive anal sex with HIV-negative or serostatus-unknown partners had lower levels of anxiety compared to their counterparts who had unprotected insertive anal sex or no unprotected anal intercourse (UAI), suggesting an inverse relationship between anxiety and sexual risk.

The relationship between anxiety and depression has bearing on these analyses. From a clinical perspective, anxiety and depression are comorbid disorders that are associated with substance use (Aina & Susman, 2006). High rates of comorbid anxiety and depression have been reported among methamphetamine users (Shoptaw et al., 2003; Vic, 2007). However, it is important to differentiate correlates of depression and anxiety because differences may inform the development of substance treatment and sexual risk reduction interventions for methamphetamine-using MSM (Starr & Davila, 2008).

Based on our review of the literature, we hypothesized that higher levels of anxiety symptoms would be associated with economic disadvantage, greater amount and frequency of methamphetamine use, injection drug use, binge use of methamphetamine, more family conflict, less social support, history of sexual assault, and more risky sexual behaviors. We also hypothesized that anxiety and depressive symptoms would have differential correlates.

This descriptive study sought to identify psychosocial and behavioral correlates of anxiety symptoms in a sample of HIV-positive, methamphetamine-using MSM. A better understanding of the correlates of anxiety could lead to further refinement of sexual risk reduction interventions and substance use treatment programs targeting this population, who represent an important source of new HIV infections in the United States and several other developed countries (Shoptaw & Reback, 2007).

Methods

Sample selection

The data for these analyses derived from a sample of 245 HIV-positive, methamphetamine-using MSM who were enrolled in a sexual risk reduction intervention in San Diego, CA. The protocol encompassed five one-on-one counseling sessions and eight maintenance sessions using a group format. Eligible participants were at least 18 years old, self-identified as MSM, reported having unprotected anal sex with at least one same-sex partner during the previous two months, and reported using methamphetamine at least twice during the past two months and at least once during the past 30 days. Participants were recruited through poster and media campaigns, street outreach, referrals from community service providers, and referrals from enrolled participants.

Measures

Anxiety symptoms

Anxiety symptoms were assessed using the six-item anxiety subscale from the Brief Symptom Inventory (BSI) (Derogatis & Melisaratos, 1983). Participants were presented with a list of problems and complaints and asked the following question: “During the past week, including today, how much were you distressed by (e.g., nervousness or shakiness inside, suddenly feeling scared for no reason, feeling tense or keyed up).” Each item was rated on a five-point scale ranging from 0 (not at all) to 5 (extremely). The BSI anxiety subscale has been shown to have good internal-consistency reliability, test-retest reliability, and construct validity (Galdon et al., 2008; Goldman et al., 2000). Cronbach's alpha was 0.87.

Depressive symptoms

Depressive symptoms were assessed using the Beck Depression Inventory-II (BDI-II) (Beck, Steer, & Brown, 1996). The BDI-II consists of 21 items that assess depressive symptoms experienced over the previous two weeks. Each item has four graded statements that are ordered (0–3) to show increasing depressive symptoms. Summary scores ranged from 0 to 63. Cronbach's alpha was 0.92.

Methamphetamine use variables

Methamphetamine use was measured as the number of grams of methamphetamine consumed in the past 30 days. The frequency of methamphetamine use was measured as the number of days on which the participant used methamphetamine in the past 30 days. Injection use of methamphetamine in the past two months was represented by a dichotomous variable (1 = yes, 0 = no).

Family conflict

Family conflict was measured using five items adapted from the family conflict scales developed by Semple et al. (1997). Participants were asked to indicate how much conflict they had had with anyone in their families in the past year because (e.g., they don't accept you for who you are; they are critical of your lifestyle). Response categories ranged from 0 (no disagreement) to 3 (quite a bit of disagreement). Cronbach's alpha was 0.88. Mean scores were used in these analyses.

Emotional support

Perceived emotional support was measured using a seven-item scale that assesses the availability of family members and friends who are perceived as caring, trustworthy, uplifting, and as confidants (e.g., “The people close to you let you know they care about you”) (Pearlin, Mullan, Semple, & Skaff, 1990). Items were rated on a four-point scale ranging from 1 (strongly disagree) to 4 (strongly agree). Cronbach's alpha was 0.91. Mean scores were calculated.

Sexual abuse

Participants were asked a single question regarding lifetime experience of sexual abuse (“Have you ever been forced or coerced to have sex against your will?”). A dichotomous response category was used (Yes = 1, No = 0).

Sexual risk behaviors

Our research group developed three scales that we used to assess sexual risk behaviors associated with methamphetamine use. The first assessed vigorous sexual practices: “When I'm high on meth: a) I have sex until my genitals are raw, sore, or bleeding; b) sex is so vigorous that condoms sometimes break or come off; and c) oral sex is so intense that my partner sometimes gets cuts and tears in his/her mouth or penis” (alpha = 0.72). The second scale captured participants' pursuit of partners at risky sexual venues: “When I'm high on meth: a) I seek out risky sexual partners (e.g., anonymous partners); b) I go to places where I know I can get sex; and c) I go cruising for sex partners at bookstores, parks or bars” (alpha = 0.79). The third scale assessed risky sexual behaviors: “When I'm high on meth: a) I have sex for hours and hours; b) I have more anal sex without a condom; and c) I will try any sexual activity” (alpha = 0.69). All items were rated on a four-point scale (1 = strongly disagree, 4 = strongly agree). Mean scores were used in these analyses.

Socio-economic and health variables

Employment and housing status were coded as dichotomous variables (employed = 1, not employed = 0; homeless = 1; other living arrangement = 0). The experience of HIV symptoms was assessed using a single item (“In the past two months, have you experienced any HIV-related symptoms?”). Responses were coded 1 = yes and 0 = no.

Statistical analysis

A multiple regression analysis was performed with anxiety symptom scores as the dependent variable. Independent variables (IVs) were entered into the regression model as a single block of variables. IVs included employment status, housing status, HIV symptoms, amount and frequency of methamphetamine use, injection drug use, family conflict, emotional support, lifetime sexual abuse, risky sexual behaviors, vigorous sexual behaviors, and seeking partners at risky venues. To test our hypothesis that anxiety and depressive symptoms have differential correlates, we also performed the above regression with depressive symptoms as the dependent variable. Diagnostic tests to detect multicollinearity did not reveal any problems with high correlations among predictor variables.

Results

Sample description

By design, all participants were male. The sample was predominantly Caucasian (58.9%), gay-identified (80.2%), never married (83.1%), living with another adult in a non-sexual relationship or living alone (55.3%), unemployed (82.4%), with a two-year degree or some college (44.9%), and an income of less than or equal to $19,999 per year (85.8%). The average age of participants was 39.8 years (SD = 7.7, median = 40.0, Range 18–61). The average score on the BSI anxiety subscale was 12.0 (SD = 5.3, median = 11.0, Range 5–30). No participants reported active treatment for anxiety symptoms. Data on socio-demographic characteristics, substance use variables, psychosocial factors, and sexual risk behaviors are presented in Table 1.

Table 1.

Sample characteristics of HIV-positive methamphetamine-using MSM (N=245)

Variable	Percentage (%)	Mean (SD)
Sexual orientation
Gay or homosexual	80.6
Bisexual	17.6
Not sure	1.6
Ethnicity
Caucasian	56.3
African American	22.9
Latino	12.7
Other	8.1
Education
Less than high school	12.2
High school or equivalent	24.9
Two-year degree or some college	44.9
Four-year college degree	11.0
Graduate or advanced degree	6.9
Marital status
Never married	84.1
Married	0.8
Separated	4.5
Divorced	10.6
Living arrangement
With same sex spouse or steady	12.7
With opposite sex spouse or steady	0.8
With other adults who are not sexual partners	27.1
Alone	28.2
Homeless	13.5
Other	17.6
Income
Less than $10,000	47.8
$10,000–$19,999	38.0
$20,000–$29,999	5.3
$30,000–$39,999	3.3
$40,000–$49,999	2.0
$50,000 or more	3.7
Employed	17.6
Experienced HIV symptoms in the past 2 months	40.7
Substance use variables
Number of years of methamphetamine use		14.0 (8.7)
Number grams of methamphetamine used in past 30 days		11.6 (27.7)
Number of days used methamphetamine in past 30 days		11.4 (9.0)
Injected methamphetamine or another drug in the past 2 months	40.0
Psychosocial factors
Family conflict		1.9 (0.89)
Emotional support		3.4 (0.62)
Experience sexual abuse, ever	49.6
Sexual risk variables
Vigorous sexual practices		2.1 (0.89)
Pursuit of partners at risky sexual venues when “high” on methamphetamine		2.9 (0.96)
Risky sexual behaviors when “high” on methamphetamine		2.9 (0.81)

Multiple regression analysis

Correlations among independent and dependent variables are shown in Table 2. In the model with anxiety defined as the dependent variable, six IVs were statistically significant. Housing status was positively related to anxiety symptoms, indicating that homeless participants had higher anxiety symptom scores compared to those who reported other living arrangements. Participants who reported HIV symptoms in the past two months and those who injected methamphetamine or another drug in the past two months also had higher anxiety symptom scores. Lifetime experience of sexual abuse, engaging in risky sexual behaviors, and seeking partners at risky venues when “high” on methamphetamine were also associated with more anxiety symptoms. Employment status, frequency and amount of methamphetamine used, family conflict, emotional support, and engaging in vigorous sex when “high” on methamphetamine were not associated with anxiety symptoms (see Table 3).

Table 2.

Pearson correlations for dependent and independent variables in regression models

	Anxiety	Depression	Number of days used MAa	Number grams of MAa used	Injection user of MAa (Y/N)b	Family conflict	Lifetime sexual abuse (Y/N)b	Seek risky partners when “high”	Go to risky venues when “high”	Have vigorous sex when “high”	Homeless (Y/N)b	Employed (Y/N)b	HIV symptoms In past 2 months (Y/N)b	Emotional support
Anxiety	1.00	.70***	.05	.06	.24***	.23***	.25***	.30***	.32***	.27***	.25***	−.16*	.26***	−.14*
Depression		1.00	.15*	.03	.21***	.20***	.19**	.25***	.19**	.26***	.31***	−.09	.20**	−.18**
Number of days used MAa			1.00	.35***	.17**	.05	.02	.13*	.13*	.13*	.33***	−.03	−.01	−.05
Number grams of MAa used				1.00	.21***	.01	.02	.13*	.15*	.14*	.12	.02	.07	.00
Injection user of MAa (Y/N)b					1.00	.03	.07	.15*	.16**	.19**	.22***	−.09	.11	−.07
Family conflict						1.00	.15*	.01	.12	.19**	.07	−.13*	.28***	−.13*
Lifetime sexual abuse (Y/N)b							1.00	.06	.12	.06	.05	−.16*	.23***	−.04
Seek risky partners when “high”								1.00	.47***	.48***	.19**	.00	−.02	.03
Go to risky venues when “high”									1.00	.45***	.18**	−.03	.02	.01
Have vigorous sex when “high”										1.00	.17**	.09	.06	−.11
Homeless (Y/N)b											1.00	−.08	.05	−.13*
Employed (Y/N)b												1.00	−.07	−.01
HIV symptoms in past 2 months (Y/N)b													1.00	−.04
Emotional support														1.00

^*p < 0.05;

^**p < .01;

^***p < 0.001;

^aMA = methamphetamine ;

^b(Y/N) (coded Yes = 1 and No = 0)

Table 3.

Anxiety symptoms regressed on socio-economic and health variables, methamphetamine use variables, psychosocial variables, and sexual risk behaviors (N=242)^a

Variable	beta	sr2
Housing status	.141 *	.016
Employment status	− .100	.009
HIV symptoms	.171 **	.026
Frequency of methamphetamine use	− .075	.004
Grams of methamphetamine usedb	− .015	.000
Injection drug use	.124 *	.013
Family conflict	.097	.008
Lifetime sexual abuse	.128 *	.015
Emotional support	− .100	.009
Engage in risky sex when “high” on methamphetamine	.168 **	.018
Seek partners at risky venues when “high” on methamphetamine	.157 *	.017
Engage in vigorous sex when “high” on methamphetamine	.047	.001
R2	.303
Multiple R	.551
Adjusted R	.267
F (df)	8.31*** (12,229)

beta = standardized regression coefficient;

^*p<.05;

^**p<.01;

^***p<.001 (2-tailed tests)

^aThree cases missing data.

^bVariable log 10 transformed

When the same regression was performed with depressive symptoms as the dependent variable, we identified two significant correlates. The experience of HIV symptoms in the past two months was positively correlated with depressive symptoms, whereas emotional support was negatively associated with this outcome.

Discussion

In our sample of HIV-positive methamphetamine users, increased anxiety symptoms were associated with homelessness, injection drug use, HIV-related symptoms, experience of lifetime sexual abuse, and risky sexual behaviors. This complex risk profile suggests that reducing levels of anxiety in the target population will be challenging and requires a multi-faceted approach.

It is not surprising that homelessness was associated with increased anxiety symptoms. In the general population, homelessness has been associated with decreased mental well-being and life satisfaction and increased risk for substance use and suicide (Eynan et al., 2002; Eyrich-Garg, Cacciola, Carise, Lynch, & McLellan, 2008). Cognitive therapy has been found effective in reducing feelings of hopelessness, stigma, and shame associated with homelessness and unemployment. Specifically, cognitive restructuring (to avoid self-labeling) and behavioral activation (daily job searching) can help reduce self-criticism, negative thoughts, and anxiety symptoms (Leahy, 2009).

Elevated levels of anxiety were also associated with HIV symptoms. Owe-Larsson, Sall, and Allgulander (2009) reported that the use of highly active anti-retroviral therapy (HAART) has been associated with decreased psychiatric morbidity, suggesting that health care providers should encourage HIV-positive methamphetamine users to adhere to HAART regimens as a possible method for reducing anxiety levels. However, the reduced HAART adherence associated with methamphetamine use (Ghaziani, 2005) suggests that HAART adherence should be promoted in methamphetamine treatment programs.

Treatment programs to reduce injection drug use have reported some success. Opioid substitution programs can effectively reduce the frequency of injection drug use and needle sharing (Metzger & Navaline, 2003). Programs that offer HIV/AIDS education, self-management of drug cravings, cognitive-behavioral strategies, and skill-building exercises have also yielded reductions in drug use (Copenhaver et al., 2006; Garfein et al., 2007). The relationship we identified between anxiety symptoms and injection drug use suggests that existing treatment programs for IDUs might also benefit from incorporating treatment for anxiety symptoms. Reyes et al. (2007) suggested that certain anxiety symptoms, such as tension and fear, could interfere with the user's ability to plan for safer sexual and injection practices as well as with condom use by elevating concerns regarding erectile dysfunction. Future studies should investigate whether treatment for anxiety symptoms in IDUs increases the effectiveness of HIV/STI prevention programs in this high-risk group (Gossop, Marsden, Stewart, & Treacy, 2002).

The association between sexual abuse and increased anxiety symptoms is consistent with previous studies of adult victims of childhood sexual abuse (CSA) (Fergusson, Boden, & Horwood, 2008; Sun et al., 2008). In our data, 70% of sexually abused participants were victimized for the first time before the age of 16. Substance abuse and anxiety disorders are common among victims of CSA. Thus, methamphetamine users who present with anxiety symptoms should be screened for past and current experiences of sexual abuse. Psychological treatments that address the adverse consequence of sexual abuse include cognitive behavioral therapy, individual and group therapy, recovery groups, hypnosis, and psychoanalysis (Walker, Holman, & Busby, 2009). Common pharmacological treatments include selective serotonin reuptake inhibitors (SSRIs) and benzodiazepines (Martin-Merino, Ruigomez, Wallander, Johansson, & Garcia-Rodriguez, 2009).

In keeping with previous studies, we found a strong correlation between anxiety and depressive symptoms. Katz et al. (1993) suggest that to disentangle depression and anxiety it is best to use a multidimensional assessment approach, including structured psychiatric interviews (e.g., Diagnostic Interview Schedule, Robins et al., 1981), self-report measures of anxiety, and observer ratings of non-verbal and expressive behaviors. Future studies should combine different sources of information to differentiate symptoms and enhance the validity of findings related to anxiety symptoms in methamphetamine users.

This study has limitations that should be taken into account when interpreting the findings. Our sample of men were volunteers in a behavioral intervention and therefore cannot be considered representative of the global population of HIV-positive, methamphetamine-using MSM. Because our study design was cross-sectional, causality and directionality in the relationship between anxiety symptoms and variables of interest cannot be determined. The present findings are also limited by the retrospective and self-report nature of the measures used to assess participants' behaviors and mood. In particular, sexual risk and substance use behaviors are subject to inaccurate recall and to reporting biases. This study is also limited by the use of a single question to measure lifetime experience of sexual abuse. Future studies should employ multi-dimensional measures that assess frequency and severity of abuse. This research also did not assess environmental stimuli or other factors that could have contributed to anxiety in the past week. Future studies should include a measure of concurrent events or circumstances that the participant might perceive as stressful (e.g., job loss). Also, the BSI anxiety subscale does not provide a clinical diagnosis of anxiety. It may be best used as a screening instrument to identify high-risk patients who should be referred for comprehensive assessment and treatment. Another limitation is that several psychosocial measures (e.g., family conflict) used time frames that differed from the period of recall for anxiety symptoms. This disparity could have resulted in an underestimation of the strength of association between some independent variables and the outcome. Last, the correlation between depressive symptoms and anxiety precluded us from using depression as a control variable in the regression model. Future studies should employ models that include depression as a control variable.

Despite these limitations, this study is one of the first to report a direct link between sexual risk behavior and increased anxiety symptoms in a sample of substance-using adults. Previous studies have focused primarily on the relationship between sexual risk behaviors and depressive disorders, although Ramrakha, Caspi, Dickson, Moffit, and Paul (2000) reported that persons with anxiety disorders were more likely to report a sexually transmitted infection (STI). The mechanism whereby negative moods affect sexual interest and behavior is largely unknown; however, it has been hypothesized that anxiety and depression can lead to sexual release as a means to manage or reduce negative arousal (Remien & Johnson, 2004). The likelihood of sexually risky behavior may also be increased by comorbid substance use and anxiety symptoms. From a clinical perspective, when elevated levels of anxiety are identified, health care professionals should screen for psychiatric symptoms and address the patient's sexual risk behaviors.