Neurological Disease in Hiv-infected Patients in the Era of Highly Active Antiretroviral Treatment: a Brazilian Experience

To study characteristics of neurological disorders in HIV/AIDS patients and their relationship to highly active antiretroviral treatment, a cross-sectional study was conducted in an infectious disease public hospital in Belo Horizonte, Brazil, between February 1999 and March 2000. Of the 417 patients enrolled, neurological disease was observed in 194 (46.5%) and a new AIDS-defining neurological event developed in 23.7% of individuals. Toxoplasmosis (42.3%), cryptococcosis meningitis (12.9%) and tuberculosis (10.8%) were the most common causes of neurological complications. The majority (79.3%) of patients were on highly active antiretroviral treatment and these individuals using HAART showed higher CD4 cell counts (p = 0.014) and presented stable neurological disease (p= 0.0001), although no difference was found with respect to the profile of neurological complications. The neurological diseases continue to be a frequent complication of HIV/AIDS and infections are still its main causes in Brazil, even in the highly active antiretroviral treatment era.

Neurological diseases occur frequently in individuals infected with human immunodeficiency virus type -1 (HIV-1) and significantly affect their functional capacity, quality of life and survival 14 18 .
Early in the AIDS epidemic, neurological manifestations were reported in 31% to 70% of all HIV1-infected patients 19 33 .Since the highly active antiretroviral therapy (HAART) has emerged however, multiple studies confirm a decrease in the incidence, as well as a change in the presentation of central nervous system (CNS) opportunistic infections and HIV associated dementia (HAD) in industrialized countries 7 11 15 16 20 21 26 30 36 37 .
The impact of HAART on the HIV-associated neurological disorders in developing countries still remains unclear.Limited access to antiretroviral therapy in addition to specific social-epidemiological conditions may produce a different picture compared with the United States and Europe 2 24 27 39 .
In Brazil, a public health program was implemented in 1996 and provides free treatment to approximately 141,000 Brazilians with HIV infection 13 34 .The effectiveness of the program has already been demonstrated in the reduction of AIDS related death 27 34 and in a significant decline of certain opportunistic infections 12 .Particularly in the field of neurological disease, however, there is limited information available.Most of the reports concerning the epidemiology of HIV/AIDS neurological disorders in Brazil were published prior to combined antiretroviral therapy 3 6 25 28 29 38 .
In this study, we determined the frequency of neurological disorders in an AIDS reference center in Brazil in the era of HAART, described the spectrum of its main etiologies and assessed the relationship between neurological diseases and antiretroviral therapy.

MATERIAL AND METHODS
This cross-sectional study was conducted in a public reference infectious disease hospital, Hospital Eduardo de Menezes (HEM), the main tertiary hospital for HIV/AIDS patients in the metropolitan area of Belo Horizonte (2.8 million inhabitants), Brazil, during the period of February, 1999 to March, 2000.
Patients.All adult HIV1-infected individuals admitted were evaluated for current neurological manifestations and/or previous AIDS-defining neurological diagnosis.Neurological manifestations were defined as the presence of any altered state of consciousness, cognitive abnormalities, motor and/or sensitivity disturbances, behavior disorders, ataxia and dizziness, cranial nervous deficit, seizures or headaches other than primary.AIDS-defining neurological diagnoses were defined according to the Centers for Disease Control, CDC/93 criteria 4 .
All medical records were reviewed for information regarding demographic variables, sexual behavior, CD4+ cell counts and HIV-1 viral load.Plasma HIV-1 RNA, expressed as copies/ml was measured by nucleic acid sequence based amplification (NASBA), with a level of detection of 80-copies/ml.Previous access to HIV related medical care prior to hospitalization and antiretroviral regime use in the 6 months prior to the evaluation period were examined.The antiretroviral treatment was classified in (1) no therapy; (2) dual therapy -patients using a combination of 2 nucleoside transcriptase reverse inhibitors; (3) HAARTpatients using three antiretroviral drugs defined as a combination of protease inhibitors (PI) or non nucleoside transcriptase reverse inhibitors (NNITR) plus two nucleoside transcriptase reverse inhibitors.
The individuals included in the study were classified into groups according to their neurological disease profile: (i) stable, i.e., patients with previous AIDS-related neurological disorders, but who were free of neurological symptoms; (ii) symptomatic, i.e., patients with signs or symptoms of neurological manifestations at admission or at any time during hospitalization.The symptomatic group was considered as new or recurrent cases of neurological disease, respectively, if the neurological diagnosis has been previous established.
The criteria for HIV-associated dementia were defined according to the AIDS Task Force of the American Academy of Neurology 1 .Primary cerebral lymphoma was diagnosed in patients with unifocal enhancing mass lesions shown by computerized tomography (CT) and positive findings in brain biopsy.Toxoplasmosis encephalitis was defined in patients with focal signs and symptoms, multifocal enhancing mass lesions in CT and improvement of the symptoms after initiation of antitoxoplasma treatment (therapy with pyrimethamine and sulfadiazine or clindamycin).Cryptococcal meningitis was considered in the presence of typical symptoms and detection of the cryptococcal antigen or direct identification of cryptococcal organisms and a positive culture in cerebrospinal fluid (CSF).Tuberculosis was presumed by typical symptoms of basal meningitis or focal sign and symptoms in addition to predominance of lymphocytes with high protein levels in CSF followed by the improvement of symptoms and CSF alterations after tuberculostatic treatment or a positive culture for Mycobacterium tuberculosis in CSF.Progressive multifocal leukoencephalopathy (PML) was defined in patients with progressive focal signs and symptoms, a decline in cognitive function and multifocal non-enhancing white matter lesions in the CT.Cytomegalovirus (CMV) encephalitis was defined as the presence of diffuse neurological signs and symptoms and CT nonspecific findings such as ventriculomegaly and periventricular enhancement associated with a positive protein chain reaction (PCR) for CMV DNA in CSF.Herpes simplex virus (HSV) and varicella zoster virus (VZV) encephalitis were considered in the presence of diffuse or focal signs and symptoms, non-enhancing focal lesions in the CT with typical pox lesions or herpes zoster.The detection of these viruses in CSF by PCR analysis was not available.Neurosyphilis diagnosis was based on the combination of reactive serological tests and abnormalities of CFS cell count or protein or a reactive VDRL-CSF with or without manifestations 5 .Bacterial meningitis was diagnosed by detection of bacteria in GRAM stain and positive latex for bacterial antigens or culture in CSF.Other diagnoses considered were psychiatric disturbances and toxic-metabolic encephalopathy.Despite extensive investigation some neurological disorders were considered unspecific when no cause could not be determined.
Data analysis.Statistical tests: chi-square test with Yates correction and the student t-test, were performed using EpiInfo 2000 version 1.12 software ceded by the Centers for Disease Control and Prevention.For all analyses p values < 0.05 were considered statistically significant.
Ethics.The Ethical Committee on Human Experimentation of the University of Minas Gerais School of Medicine and the Eduardo de Menezes Hospital approved the study.
The descriptive characteristics of those with neurological disorders can be seen in Table 1.Age ranged from 18 to 65 years old (mean age = 35.8± 0.6), 135 patients (69.6%) were male, 142 (89.3%) had a low education level (≤ 8 years) and 98 (80.3%) acquired HIV through sexual transmission.The mean T lymphocyte CD4 cell count was 93.9 X 10 6 /l (SD=110.6)and mean plasma HIV-1 RNA was 28,854 copies/mm 3 (SD=602.3).Thirty-eight (19.6%) individuals with neurological disease were diagnosed with HIV-1 infection at the hospital, while 156 (80.4%) patients presented a previous well-defined diagnosis, among which 115 (77.7%) were under HIV outpatient care prior to admission.Finally, 111 (76.1%) had been on antiretroviral therapy in the 6 months prior to the evaluation period, among which, 88 (79.3%) had been on HAART, 12 (10.8%) on dual therapy and for 11 (9.9%) individuals there was no information regarding the antiretroviral scheme used.
Infections occurred in 151 (77.8%) patients.Toxoplasmosis was the most frequent opportunistic infectious disease (42.3%), followed by cryptococcosis meningitis (12.9%) and tuberculosis (10.8%) (Table 2).HIV-associated dementia was observed in 9 (4.6%) cases in this series and primary CNS lymphoma in only one (0.5%) patient.Despite extensive neurodiagnostic and neuroimaging evaluation, no etiology was defined in 16 (8.3%) of the individuals.Symptomatic neurological HIV-infected patients presented statistically significant lower T CD4 cell counts (≤ 100 cells/mm 3 ) (p= 0.045), were less likely to have used previous HIV outpatient care (p=0.005) or to have used antiretroviral treatment (p=0.001), compared to stable patients.There were no differences which respect to gender, age, educational level and viral load between the two groups (Table 3).
The variables that were associated with being on HAART were lymphocyte T CD4+ cell count ( >200 cells/mm 3 ) (p=0.014) and stable neurological profile (p=0.0001)(Table 4).There were no differences between the antiretroviral schemes with respect to viral load, overall AIDS-defining neurological disease or overall opportunistic infection.However, the numbers were too small for proper statistical assessment.

DISCUSSION
This is the first study on neurological disorders in the HAART era in Brazil.The most important finding of the study was that neurological disorders continue to be a frequent complication in HIV1-infected patients.The observed proportion (46.5%) of cases was very similar to publications prior to 1996.In two clinical series between 1989 and 1991, Vecino et al 38 found a prevalence of 60.4% for neurological disease and Puccioni-Sohler et al 28 demonstrated a frequency of 26% of neurological complications in two distinct Brazilian university hospitals.Our results were also in agreement to that found a prevalence of 42.7% for neurological manifestations in São Paulo 9 .
Brazil is the first developing country to have implemented a large-scale universal antiretroviral distribution program and approximately 141,000 patients now receive antiretroviral treatment through the public health system.This study describes the overall characteristics of neurological disorders in HIV after universal access to HAART in Brazil.In this context, the current findings may indicate an apparent contradiction.Considering that 76% of neurological patients were on antiretroviral treatment and nearly 80% of them on HAART, a lower prevalence of neurological disorders would be expected.Although a complete explanation is difficult, the results presented here could be partially biased by the severity of the patients' disease conditions.Data were collected from a hospitalized population, presumably more immunosuppressed than outpatients.As demonstrated by the low mean T CD4+ cell count (93.9 x 10 6 cell/l), these patients indeed presented end-stage AIDS and were expected to develop more opportunistic complications, including CNS diseases.The possibility of partial or deferred immune reconstitution in response to HAART should also be considered.The drug regimen and the onset the antiretroviral therapy were also not investigated.It is possible that patients who had recently started HAART may have been subject to a higher risk of a neurological event.
Other factors such as a poor adherence to HAART, a less potent regimen, a failing regimen, or a recent rescue scheme may have contributed to the increased risk of neurological complications.
We characterized the profile of patients who were developing neurological disease in the era of combined antiretroviral treatment.Individuals with advanced immunodeficiency, T CD4+ cell counts < 100/mm 3 (p = 0.045), those without previous HIV outpatient care (p=0.005) and patients not receiving antiretroviral therapy (p=0,001) showed a statistically significant probability for the manifestation of neurological symptomatic disorders.The percentage of patients on HAART was also significantly higher in the group of neurological stable patients (p=0.0001),compared to the symptomatic patients.This same profile has been observed in the developed world, where central nervous system diseases still occur, particularly in individuals who do not take antiretroviral drugs, but even in those who do take theses agents 8 15 20 22 31 .
Infections were the most common cause of neurological disease and occurred in 151 (77.8%) patients, which suggests that the spectrum of neurological disease etiologies has not changed in Brazil.Our results are consistent with Chimelli at al 6 in a pre-HAART report on 252 autopsy cases who observed a 65.4% incidence of opportunistic infections and Fragoso et al 9 who found an 80% rate of opportunistic neurological disease in an early-HAART Brazilian series.It is probable that geographic and socioeconomic factors interfere with the evolution of the HIV-1 infection, possibly by enhancing the risk of exposure to infectious agents.Reports from India 17 32 and Mexico 10 23 confirm the predominance of infectious neurological disorders over HIV dementia and neoplasic disease.
Toxoplasmosis was still the most common opportunistic infection in the central nervous system (42.3%) and cryptococcal meningitis was the second (12.9 %).The present study could not confirm the low frequencies of toxoplasmosis (2.2% to 8%) and of cryptococcosis (0.8% to 4.9%) documented by Masliah et al 21 Sacktor et al 31 , Mascke et al 20 , Jellinger et al 15 , Neuenburg et al 26 , and Gray et al 11 in the HAART era in the developed world.Tuberculosis was the third cause (10.8%) of opportunistic neurological infection in this study in agreement with statistics from tropical areas and in developing countries 17 32 .The frequency of HIV associated dementia and primary CNS lymphoma were very low in this Brazilian series, probably misdiagnosed because the concomitance of opportunistic infections and the limited access to cerebral biopsy and modern imaging.
We believe there are several limitations to the current study.Only HIV-infected inpatients were considered.It is unknown whether similar results would be obtained in other groups, such as outpatient care or with a higher T CD4+ cell count and higher adherence.Data related to antiretroviral regimens and compliance was not always available.The low specificity of neurological signs and symptoms and the limited access to diagnostic methods could have contributed to the high prevalence of presumptive opportunistic infections and may have impaired the recognition of non-opportunistic AIDS-defining neurological disorders.
A larger, multicenter study to analyze the temporal trends of the neurological disease in HIV/AIDS patients in Brazil and its changing profile is warranted.In addition, earlier HIV testing, diagnosis and proper follow-up, including primary prophylaxis for opportunistic infections and HAART, should be emphasized as public health policies for the treatment and prevention of AIDS in Brazil and other developing countries.