Prevalence and factors associated with the generalized anxiety disorder among people living with HIV/AIDS in Brazilian outpatient clinics

Introduction: The Generalized Anxiety Disorder (GAD) is a chronic worrying disorder, which causes social and occupational impairment. Objective: The aim of this study was to estimate the prevalence of GAD among people living with HIV/AIDS (PLWHA) and to determine the associated socio-demographic and clinical factors. Methodology: We conducted a cross-sectional study between September 2014 and April 2015 at four HIV outpatient clinics in Fortaleza, Northeastern Brazil. We surveyed 257 individuals living with HIV/AIDS, aged > 18-years, of both sexes. Four modalities of instruments were used to collect the data: a socio-demographic questionnaire; the Mini International Neuropsychiatric Interview; the Hamilton’s Anxiety Scale; and chart data. Results: GAD prevalence was 14%. Female sex, CD4 200-500 versus < 200, CD4 > 500 versus < 200 and being homeless were associated in the multivariate model. Conclusion: The study demonstrates the necessity of proactively addressing mental health as well as physical health issues in HIV+ individuals.


INTRODUCTION
The infection by the Human Immunodeficiency Virus (HIV) affects approximately 37 million people worldwide, 1 and 734,000 people in Brazil. 2 HIV incidence and mortality have declined in most parts of the world.However, HIV incidence in Brazil has increased. 3][12] Other factors accentuate the importance of studying GAD, such as: it is frequently combined with other mental disorders and increases the likelihood of first onset of mood disorders, panic disorder and substance abuse disorders; 13,14 it represents an independent risk factor for coronary morbidity, and causes high economic burden. 14e aim of this study was to estimate the prevalence of Generalized Anxiety Disorders (GAD) among PLWHA in a major Brazilian city, and to determine associations with socio-demographic and clinical factors.

METHOD
We conducted a cross-sectional study between September 2014 and April 2015, in four public HIV clinics in Fortaleza, the fifth most populous city in Brazil.A convenience sample of HIVinfected patients awaiting clinic visits were invited to participate.The purpose and methodology of the survey were explained to all candidates, and written consent requested.Individuals of both sexes, 18-year-old and over, diagnosed with HIV, whether or not in antiretroviral treatment, were included.Individuals with dementia, drunk or otherwise incapacitated were excluded.The Mini-Mental State Examination (MMSE) 15 was used to assess cognitive impairment.There was no language barrier.Four modalities of data collection instruments were used: (a) A Socio-demographic questionnaire elicited information on sex, age, marital status, income, time since diagnosis of HIV, and use of antiretroviral medication; (b) A section of the Mini International Neuropsychiatric Interview -MINI 16 was used to assess GAD; (c) For patients diagnosed with GAD, we used Hamilton's Anxiety Scale (HAM-A); 17 (d) We reviewed chart data for information on clinical aspects, history of opportunistic infections, antiretroviral treatment (ART), and viral load and CD4.
MINI is a structured questionnaire to diagnose mental disorders as encoded by the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) and the International Classification of Diseases and Related Health Problems (10th ed.; ICD-10; World Health Organization, 1992).This instrument contains 16 modules identified by letters of the alphabet.It has been previously used to assess mental disorders in HIV + individuals in other studies. 18,19In this study, module O (Generalized Anxiety Disorder -GAD) was used.We attempted to exclude any anxiety symptom related to other mental disorders, such as the fear of panic disorder or public humiliation.HAM-A is an instrument with rating scales developed to measure the severity of anxiety symptoms, and is widely used in both clinical and research settings.We interpreted the scores following Matza, Morlock, Sexton, Malley & Feltner. 20he cases of GAD were classified into three groups: mild = 8 to 14; moderate = 15 to 23; and severe > 24.
Data were entered using EpiInfo (version 3.5.3)and analyzed with SPSS (version 11) utilizing chi-squared (χ2) and Fisher's exact tests.We applied two-tailed tests with α = 0.05.Variables significant at p < 0.05 were included in the multivariate model.We used a logistic regression to determine the factors predicting the occurrence of mental disorder.Confidence interval was set to 95% at the 0.05 level of significance.
Our research did not impose any risk to the studied patients.Participants were consented and we offered counseling to patients diagnosed with a mental disorder.The study was approved by the Research Ethics Committee of the Hospital São José de Doenças Infecciosas (Protocol 670.113).

RESULTS
We surveyed 264 individuals.Seven declined to participate in the study.The majority of the surveyed PLWHA fell within the following parameters: between 30 and 59 years-old (80.5%), male (61.9%), single (36.2%) and monthly income of up to 2 minimum wages (54.1%), around US$ 480,00.The GAD prevalence was 14% (Table 1).Concerning the severity of GAD, 7.4% were considered mild cases, while 22.2% and 7.4% were considered moderate and severe, respectively.
Among women, the prevalence was of 20.4%, greater than for men (p = 0.020).GAD prevalence was also associated with lower income (p = 0.020).The prevalence was significantly higher in those who had been homeless (p = 0.014).In addition, GAD prevalence was associated to those cases with a family history of GAD and with those patients who had never presented CD4 < 200 cels/dL (Table 2).Most participants with GAD (68,8%) were not receiving proper treatment for the disorder by the time of the interview.

DISCUSSION
The GAD prevalence found in our study (14%) is much greater than that of the general population (3.7%). 21Other studies found a GAD prevalence of 2.2%, 22 1.4% 11 and 3.6%, 23 probably owing to methodological issues, such as the use of a semi-structured instrument or a smaller sample size and cultural/regional differences.
Besides the association found between GAD and lower incomes, 66.7% of the individuals were unemployed.This finding highlights how income and work related issues might influence the expression of GAD, as revealed by other studies. 10,24Moreover, the fact that we detected an association between GAD and having been a homeless supports the relationship between poor socioeconomic conditions, stigma and mental illness. 96][27] In our study most of the patients were enrolled in ART.This might be due to the fact that the sample was constituted by patients who were attending medical appointments, which, we might suppose, would more probably follow the treatment prescribed than those who didn't attend appointments.Also, this might reflect better access to ART by general population.
While some studies show higher GAD prevalence in individuals with lower CD4's 24 or did not show any significant association, 28 in our study higher CD4 values were correlated with the GAD prevalence.Such an apparently controversial finding may follow the same rationale explained above: worried patients may have initiated ART earlier, reflecting higher CD4s.Here we draw a parallel between the "worried well" and the "worried ill".
Some antiretroviral drugs, such as efavirenz, are known to cause psychiatric side effects like depression, insomnia and anxiety. 29One might think that the association between higher CD4 and GAD would reflect ART adherence, especially efavirenz.However, we found no association between efavirenz use and GAD (Table 2).
Few studies assessed the severity of GAD on PLWHA.Els et al. 30 identified HAM-A scores above 20 in 31% of the sample.
In our study, higher scores were present in most patients OR: Odds ratio; CI: confidence interval.
(70.4% indicate severe anxiety).These are worrying facts, since most of the participants with GAD (68.8%) were not prescribed a specific treatment for the disorder.
Our study has an important limitation.The sample is a convenience sample, drawn from a clinic setting, and may not represent the majority of PLWHA.As mentioned, GAD and clinic attendance may not be independent, and may have greatly influenced our findings.The profile of our study patients corresponds to that of those who regularly frequent medical services.In general, available laboratory tests demonstrate adherence to treatment, which may not correspond to the reality of a great number of HIV-infected patients with mental disorders.While curiously we show an association between GAD and higher CD4 counts, this does not obviate the suffering and social and economic effects of GAD.Certainly, this relationship deserves further investigation.
We conclude that GAD is present in a large number of HIV+ patients.Given that, the condition should be addressed in all clinical services that address persons living with HIV/AIDS.

Table 1 .
Socio-demographic characteristics of HIV-infected patients from outpatient clinics of Fortaleza, Brazil.

Table 2 .
Factors associated to GAD among HIV-infected patients of Fortaleza, Brazil.

Table 2 .
Factors associated to GAD among HIV-infected patients of Fortaleza, Brazil.

Table 2 .
Factors associated to GAD among HIV-infected patients of Fortaleza, Brazil.

Table 3 .
Multivariate model for the factors associated to GAD in HIV-infected patients of Fortaleza, Brazil.