Introduction

The struggle against AIDS (Acquired Immune Deficiency Syndrome) started in the 1980s and the disease has become pandemic, particularly in developing countries. Until the end of 2003, 40 million children and adults were living with HIV/AIDS worldwide. There were about 14,000 new HIV infections per day in 2003; of those, more than 95% were in low and middle income countries, and almost 2,000 occurred in children under 15 years of age [31].

Especially for youngsters, the UNAIDS/WHO report [30] emphasizes the need for preventive efforts, since recent data from several countries have demonstrated an increase in the incidence of risky behaviors, low frequency of condom use, and higher rates of sexually transmitted diseases (STDs) in this specific group.

Evidence from different sources document that young people with serious mental disorders presents high risk for HIV infection [1]. There were frequent reports of depression, anxiety, bipolar disorder, and sexual abuse trauma in samples of North-American seropositive adolescents. Many of these conditions preceded HIV infection itself, and were amplified in its impact by the positive result of the HIV test [15].

It is important to note that due to the high association between mental health problems, drug use/abuse and sex risk behaviors, it is extremely difficult to disentangle the isolated effects of mental health problems in the pathways ending into HIV infection. Thus, adolescents with psychiatric symptoms might be at a greater risk for HIV infection due to a greater prevalence of sexual relations, multiple partners and sexual abuse [22]. Ramrakha et al. [26] identified a higher probability of risky sex associated with diagnoses of mental disorders at the age of 21. Depression, drug addiction or anti-social personality increased the chances for risky sexual behavior, sexually transmitted diseases, and early sexual life in that sample. In addition, personality traits—such as impulsiveness and sensation seeking [20, 33]—may predispose adolescents to early sexual debut, drugs and other illegal behaviors. Marijuana, cocaine, and other drugs have been associated with a greater rate of sexual relations, multiple partners and low use of condoms [18]. Donenberg and Pao [11] in a recent review with a focus on child and adolescent risk factors associated with HIV/AIDS, showed their significant mental health implications.

Little is known about the relation between sexual risky behaviors and psychiatric disorders that reach their peak during adolescence with important implications for public health. Fleitlich and Goodman [12] suggest that a great share of Brazilian adolescents need special care for their high prevalence of mental disorders. Since HIV infection is spreading quickly in the young population of developing countries, it is of surprise the scarcity of research to assess the role of mental health problems for the development of HIV infection in the context of drug use patterns and sex risk behaviors in these countries. For instance, there is very little research analyzing the connections between mental health problems, drug abuse, or STDs in Brazil [27]. Thus, the main objective of this study was to investigate the association between psychiatric symptomatology and HIV infection in a sample of young people from a developing country, controlling for the impact of risky behaviors involving sex and drugs. Our main hypothesis was that HIV+ adolescents would show higher psychiatric symptomatology compared to HIV− adolescents, even when other risky situations were controlled.

Method

Sample

The sample was comprised of consecutive adolescents aging 13–20 years-old assessed in an anonymous testing site for HIV and STDs in the city of Porto Alegre during one year. Porto Alegre is the capital of the southernmost state of Brazil with a population of 1,800,000 inhabitants. The only exclusion criteria were the lack of cognitive conditions or altered conscience state that would prevent interview (assessed clinically), and involuntary attendance to the center (e.g., sent by the court). This sample is not representative of the whole population of adolescents from the city. Due to difficulties in the system of referral/assessment of this Public Health Clinic, we could not obtain the exact number of individuals who were potentially screened for the study. The Clinic’s information is that there is an universe of about one thousand adolescents (13–20 years old)/year that search for anonymous HIV tests.

The study was conducted with the adolescents’ tacit consent—only those who accepted to participate (n = 402) answered the forms. Subjects were informed of the goals of the study and filled out the questionnaires anonymously. The study was approved by the Ethics Committee and Institutional Review Board at the Hospital de Clínicas de Porto Alegre (approved as an IRB by the Office for Human Research Protections, United States of America-IRB 00000921).

Design

This is a cross-sectional study. The study factor was the presence of psychiatric symptoms. The outcome was HIV status. Potential confounders included demographics (age, gender, marital status, employment, family income, and schooling) and sex risk behaviors for HIV and drug use patterns in the last 30 days and throughout life.

Procedures and instruments

Data were collected by trained interviewers who were blind for the respondent’s serostatus. Standard pre-test counseling data was performed by the center’s permanent staff. Adolescents answered questionnaires individually. Interviewers were trained in a pilot study with data collection of 19 cases, which were excluded from final analyses. These pilot cases were used to calibrate the standards for data collection and the development of the database. Instruments included:

  1. (a)

    SCL-90-R- The Symptom Checklist-90-R [8, 9, 16] is a standard instrument that allows for detection of psychiatric dimensions of symptomatology in the week prior to the interview in the following areas: somatization, obsessive compulsive, interpersonal, depression, anxiety, hostility, phobia, paranoid ideation, and psychosis. It is a screening instrument for global psychological discomfort, discriminating between cases and non-cases [9].

  2. (b)

    RAB—Risk Assessment Battery [21]. The RAB is a self-applied questionnaire translated into Brazilian Portuguese [24, 25]. It assesses drug and sexual behaviors throughout life, in the previous six months and during the month prior to data collection, as well as basic demographics data. The instrument comprises 38 questions. Questions about the age of sexual initiation, pregnancy, and/or previous abortions were added to the RAB.

  3. (c)

    ELISA (Enzyme Linked Immunosorbent Assay) test and confirmatory tests to evaluate HIV serology. All blood samples were tested using two ELISA essays with different antigens. The exams with a positive, indeterminate or discordant result were confirmed using the Indirect Immunofluorescence technique—IFI. Indeterminate Immunofluorescences were confirmed by the Western Blot test.

The database was subjected to systematic internal analyses, when questionnaires were individually checked, retyped and compared by an external observer.

Data analyses

Quantitative data were described by means and standard deviations, and qualitative data by absolute and percentile frequency. Odds ratios (OR) were calculated with their respective confidence intervals of 95% and their significance was determined by the chi-square test. Fisher’s exact test was used for low frequencies. Relations among quantitative data were evaluated by the Student’s t test and the magnitude of associations was determined by the effect size. All tests were two-tailed. The level of significance was set at 5% and the data were analyzed using SPSS version 11.0.

After bivariate analyses, we performed multiple logistic regression analyses. To avoid excessive testing (Type I Error) and collinearity, we entered three composites: psychiatric symptomatology (overall score in the SCL-90-R), sexrisk (based on all individual variables along lifetime, for both genders, that were associated with positive HIV status in bivariate analyses for a P < 0.2: sexual relation with male partners without condom + sexual relation with an injection drug users (IDU) partner + money in exchange for sexual relation + sexual debut < 12 years old + never wore condom in sexual relation) and drugrisk (based on the following individual variables: illicit drug use+ been in drug-using/drug-selling places; those presenting a P < 0.2 in bivariate analyses). Two demographic variables (low schooling and family income) were associated to positive HIV status for a P < 0.2. Since they were highly associated, only low schooling was kept in these multivariate analyses to avoid collinearity.

Results

The sample comprised 402 youngsters who filled the instruments. About 14 cases were lost, from whom no HIV test results could be obtained. The overall HIV-positive status of this sample was 6.2%, based on 388 respondents. Mostly, the sample was comprised of young, predominantly female subjects, with little schooling (50.9% did not complete elementary school) and income. Although a positive serostatus was significantly associated with low schooling (P = 0.03; OR = 7.8, CI95% = 1.1–59.3), no other significant difference was detected between positive and negative HIV adolescents regarding demographic variables (gender, age, family income, marital status, and job status).

With regard to psychiatric symptomatology, seropositive adolescents showed significantly higher scores in all dimensions of the SCL-90-R (see Table 1). The overall SCL-90-R score was significantly higher in positive HIV adolescents (P < 0.01; Effect size = 0.87).

Table 1 Psychiatric Symptomatology and its association with HIV serostatus
Full size table

Regarding sexual risk factors throughout life, positive HIV status was significantly associated with early sexual initiation (before 12 years of age) (P = 0.02), and sexual intercourse with a male partner without condoms (P = 0.03). Other variables not significantly associated with HIV seropositivity, but included in the SEX composite in the multivariate analyses (P < 0.2) were: sexual relation with an injection drug user (IDU) partner, money in exchange for sexual relation, and never wore condom in sexual relation. Infrequent condom use was high for both groups (Table 2). As for sex risk practices in the six months prior to the study, adolescents who reported having received money for sex showed a higher association with seropositivity (P < 0.01), the same happening with those who had sexual relations with a possible seropositive partner (P < 0.001).

Table 2 Sex risk practices along lifetime and its association with HIV seropositivity
Full size table

The overall prevalence of drug use (Table 3) in the month prior to data collection was high in both groups. The most frequently reported drugs used were alcohol, marijuana, snorted cocaine, and inhalants. HIV positives used significantly more illicit drugs (P = 0.01). Specifically, recent marijuana (P < 0.05), amphetamines (P < 0.01), and solvents uses (P < 0.001) were significantly higher in positive HIV adolescents. Attendance to drug-using/drug-selling places showed statistically significant association with seropositive status (P < 0.04).

Table 3 Risk behavior through drug use previous to the sampling and its association with HIV seropositivity
Full size table

Findings from multivariate logistic regression analyses can be found in Table 4. When schooling and the three composites (psychiatric symptomatology, sexrisk and drugrisk) were included in the final model, only the SCL-90 overall score was significantly associated with seropositivity (P = 0.032; OR = 1.88, CI 95% = 1.06–3.34). The SEX-RISK composite reached borderline significance (P = 0.06; OR = 1.63, CI 95% = 0.98–2.70).

Table 4 Multiple logistic regression analyses for composite risk factors and HIV seropositivity
Full size table

Discussion

We were able to document that psychiatric symptomatology was significantly associated with HIV seropositivity in a sample of adolescents from a developing country even adjusting for other relevant risk behaviors. To the best of our knowledge, this is the first study comparing adolescent HIV serostatus and psychiatric symptoms in the context of risk behaviors in a developing country with high rates of HIV infection.

Voluntary HIV testing is uncommon among young people, and UNAIDS [30] has reported adult seroprevalence of 0.6% in Latin America. The high seroprevalence found in this sample may be due to the fact that a significant fraction of this group had a greater perception of risk than the general population, which may have motivated them to look for a test center. It seems that they represent a highly selected group of young people that voluntarily choose to find out their serologic status, maybe motivated by media.

Although the preponderance of HIV infection in adolescent females did not reach statistical significance in our study, data published by the Brazilian Ministry of Health show an increase in the proportion of infected adolescent females compared to male adolescents, from 0.6 in 2001 to 0.7 in 2004 [3, 5], pointing to younger women as the group where the epidemic most spreads in the country.

It is important to note that HIV infection and psychiatric symptoms and/or disorders have a complex relationship. In samples of adults, there are several evidences documenting the link between HIV infection and psychiatric disorders. For example, Kelly et al. [17] found elevated lifetime rates of major depression in HIV-positive men in a cross-sectional study comparing HIV-negative and HIV-positive homosexual/bisexual men in Australia. In the HIV-positive men, psychiatric disorder was significantly associated with the presence of lifetime psychiatric disorder prior to HIV infection diagnosis. In adults too, Thompson et al. [29] provided evidence that patients with serious mental illness had higher rates of participation in risk behavior for HIV infection than those in the general community. However, data from samples of adolescents are scarce. Severely mentally ill youths are at elevated risk for human immunodeficiency virus infection, but little is known either about acquired immunodeficiency syndrome (AIDS) risks behaviors in adolescents who seek outpatient mental health services or the links between psychiatric problems and particular high-risk behaviors [10]. Our findings suggesting higher psychiatric symptomatology in positive HIV adolescents are in agreement with previous report from the literature [1, 6, 14, 22]. In addition, the association was kept even after adjusting for the impact of other relevant risk factors such as sex risk behaviors and drug use. Moreover, this association was verified in a complete different culture than those previously assessed.

Sex risk behaviors were marginally associated to positive HIV status in multivariate analyses. Among sex risk behaviors, early sexual initiation seems to be a relevant factor for HIV infection. Most teenagers reported sexual debut between 12 and 16 years of age and those who started before 12 were at even greater risk for HIV infection. Delaying sexual initiation has an extra advantage of protecting adolescents from STDs and unwanted pregnancies. Whaley [32] suggests the joint approach to both AIDS and unwanted pregnancy would favor care for female adolescents with regard to these aspects. In addition, the infrequent use of condoms and carelessness in preventing pregnancy detected in this study is in agreement with reports from other studies [2, 14]. More than 80% of the adolescents reported sex without condoms. Despite nationwide media campaigns, the information is not used to guide adolescents’ conduct even in those who perceive themselves as belonging to a high risk group. It is possible to speculate that lack of schooling, altered perception generated by drug use and/or psychiatric symptoms may make these subjects feel invulnerable. Attention to HIV education and prevention in adolescents has been inappropriately scant; strategies to encourage safer behavior are urgently needed.

Although non-injectable drug use was associated with positive HIV in bivariate analyses, the significance disappears in the context of psychiatric symptomatology, low schooling and sex risk behaviors. It is important to note that marijuana use in the month previous to data collection was lower in this sample than those reported in other studies with Brazilian adolescents [13]. Since other investigators reported that mental health disorders like depression [4, 7], bipolar disorder, conduct disorder and ADHD might be risk factors for substance use disorders [19, 28], drug abuse might be a intermediate factor in the pathway between mental health problems and HIV infection.

Limitations

The findings of this study should be understood in the context of some methodological limitations. First, the generalization of our findings to other geographic or socioeconomic groups is restricted due to the low schooling and income of our sample. Second, for logistic reasons it was not possible to test the subjects for IQ. Third, the sample was collected at a free testing site. This fact could have generated a greater rate of seropositivity as opposed to non-clinical samples. Fourth, the screening instrument chosen for the assessment of psychiatric symptoms (SLC-90-R) does not cover all psychiatric disorders, although it is a broad-band instrument [8, 9, 16]. As stated by Myers and Winters [23], most scales do not provide all the psychometric data desired in selecting the best scale. Moreover, it is important to note that all SLC-90-R psychiatric dimensions were significantly associated with the HIV status. Fifth, the small sample size of positive HIV adolescents might have prevented the detection of some clinically relevant associations. Finally, only an association (not causal relationship) between psychiatric symptoms and HIV status might be inferred from this study due to the cross-sectional design.

Clinical implications

Health professionals dealing with policies for prevention of HIV infection in young people in developing countries are quite familiar with social and environmental risk factors, but much less trained in the recognition of the impact of mental health problems on seropositivity. Thus, our findings suggesting an association between psychiatric symptoms and positive HIV infection indicate that among young people asking for HIV testing in Brazil, seropositivity is associated with psychological symptoms and that screening for the latter would therefore be appropriate in this context.

In addition, longitudinal studies are needed to identify specific roles of psychiatric symptoms, sex risk behavior, and drug use in the pathway ending in HIV infection during adolescence, especially in developing countries.

Disclosure: Dr. Rohde is on the speakers’ bureau or is a consultant for Bristol-Myers Squibb, Eli-Lilly, Janssen-Cilag, and Novartis. Other authors have no potential conflict of interest.