HTLV-1 and HTLV-2 Infections in HIV-Infected Individuals From Santos, Brazil: Seroprevalence and Risk Factors

Because HTLV-I, HTLV-2, and HIV share identical modes of transmission, simultaneous or subsequent infections with these retroviruses are to be expected. The population of Santos, the largest port in Latin America, includes large numbers of female commercial sex workers and intravenous drug users, presumably having been exposed to retroviral infection. To evaluate the seroprevalence of HTLV infection and their associated risk factors, a cross-sectional survey was carried out in 499 HIV-infected individuals from Santos, Brazil. HTLV testing consisted of enzyme immunoassays for serologic screening and confirmatory Western blot testing. Overall HTLV-I and HTLV-2 seroprevalences were 6.0% (95% confidence interval [CI], 3.9-8.1) and 7.4% (95% CI, 5.1-9.7), respectively. Multivariate logistic regression for statistical analysis revealed HTLV-I infection to be independently associated with: intravenous drug use (IDU) (odds ratio [OR]. 2.99; 95% CI, 1.09-8.20), seropositivity to hepatitis C virus (HCV) (OR, 3.03; 95% CI, 1.02-9.01) and < 3 years of education (OR, 4.73; 95% CI, 1.56-14.41). HTLV-2 infection was associated with: IDU (OR, 3.22; 95% CI, 1.33-7.84), HCV seropositivity (OR, 5.40; 95% CI, 1.86-15.66) and nonwhite race (OR, 3.32; 95% CI, 1.58-7.00). Results indicate that HIV-infected individuals living in Santos are at similarly high risk of being exposed to HTLV-1 and HTLV-2. IDU constitutes the main risk factor for HTLV acquisition in this population, and there is no significant risk associated with sexual practice.

HTLV-1 and HTLV-2 are members of the Retroviridae family, which also includes HIV ⁽¹⁾. Even though HTLV-1 and HTLV-2 infections are found worldwide, certain geographic regions bear higher endemic levels: HTLV-1 is found in southern Japan, markedly in Kyushu and the Okinawa islands, the Caribbean basin and South America, sub-Saharan Africa, and Melanesia ^(2-4). HTLV-2 is mainly found among native indigenous populations of North, Central, and South America ^(5-7) and in African pygmy tribes from former Zaire and from Cameroon ⁽⁸⁾. In the United States, HTLV-2 infection is particularly prevalent among intravenous drug users (IVDUs) ⁽⁹⁾. Given that HTLV-1, HTLV-2, and HIV possess identical modes of horizontal and vertical transmission, retroviral coinfections may be diagnosed in populations that are highly exposed to common risk factors. In fact, HTLV-1 and HTLV-2 coinfections are frequently seen among HIV-infected individuals in Brazil and elsewhere ^(10-13). As the largest port in Latin America, Santos hosts over 3,000 ships every year, and about 2,000 trucks every day, which represents a significant and continuous population flow from several locations both in Brazil and abroad. Moreover, because of poverty, Santos presents a large scope for prostitution and, as part of the international cocaine route, it has a large population of IVDUs ⁽¹⁴⁾. Epidemiologic data obtained from the Brazilian Ministry of Health not only indicate that Santos has one of the highest AIDS prevalences in the country, but also that HIV transmission in the city mainly occurs through the use of intravenous drugs ⁽¹⁵⁾. In this context, a seroprevalence survey of HTLV infections in such an environment may be one means of characterizing the dynamics of retroviral infections in a population that combines all necessary elements for their acquisition.

METHODS

Selection of Study Subjects

The study was carried out in a cohort of HIV-infected individuals, assisted at Santos AIDS Reference Center. The institution provides outpatient medical care to HIV-infected people and is the only public health center with these features in the city. In it is thus concentrated demand from the entire local public system demand and also some neighboring towns. Patients were enrolled either at admission to the center or during regular medical visits and were included after having signed an informed consent. A seroepidemiologic survey was then carried out in a cross-sectional design. By means of a questionnaire, sociodemographic data were collected, as well as information on sexual practice, history of previous sexually transmitted diseases (STDs), intravenous drug use (IDU), earlier blood transfusions, and breast-feeding. Additional relevant clinical and laboratory data, including previous history of opportunistic infections and CD4⁺ cell counts, were obtained from medical records.

Laboratory Investigation

Blood samples were screened for anti-HTLV-1/II antibodies with two immunoenzymatic assays (EIAs), based on different antigenic bases: 1) viral lysates from HTLV-1-infected and HTLV-2-infected cell lines (Hemobio HTLV I/II, Embrabio, Brazil); and 2) recombinant peptides, derived from HTLV-1 and HTLV-2 transmembrane proteins, spiked with synthetic peptides reproducing immunodominant regions of the viral envelope (Murex HTLV I+II GE80/81, Murex Diagnostics, Brazil) ⁽¹⁶⁾. Concordantly positive and discordant samples on EIA were selected for confirmatory testing. Western blot tests (WB), which allow not only confirmation, but also differentiation between HTLV-1 and HTLV-2 infections, were used (GLD HTLV BLOT 2.4, Genelabs Diagnostics, Singapore). These incorporate a recombinant peptide derived from the viral transmembrane protein (GD21) and type-specific peptides derived from the external viral envelope protein (rgp46-I and rgp46-II) ⁽¹⁷⁾. Testing was done according to manufacturers' recommendations. WB results were interpreted as follows:

Seropositive for HTLV-1: seroreactivity to both gag (p19 and/or p24) and env (GD21 and rpg46-I) antigens

Seropositive for HTLV-2: seroreactivity to gag (p19 and/or p24) and env (GD21 and rpg46-II) antigens

Seropositive for HTLV-1/II: reactivity to p19, p24 and GD21 only

Seroindeterminate: reactivity to viral antigens that does not fulfill the above mentioned criteria

Seronegative: no seroreactivity to viral antigens

To investigate existing sexually transmitted or blood-borne coinfections, blood samples were also screened for serologic markers of hepatitis B virus (HBV) (HBV surface antigen [HbsAg] and anti-HBc), hepatitis C virus (HCV) (anti-HCV) and syphilis, using standard EIA (Enzymun-test, Boehringer-Mannheim for HBV and HCV and TP-Hemagen for anti-Treponema pallidum immunoglobulin G).

Statistical Analysis

Obtained data were loaded in the Epi-Info software, version 6, produced by the U.S. Centers for Disease Control and Prevention (CDC) (Atlanta, GA, U.S.A.). HTLV-1 and HTLV-2 seropositivities were considered dependent variables. Association with independent variables was measured in univariate analysis, using the χ² test, and establishing odds ratios (ORs) and their respective 95% confidence intervals (CIs). For discrete quantitative variables, the continuity correction of Yates was applied. Multivariate analysis was carried out, based on logistic regression models, using the standard software Epidemiological Graphics, Estimation and Testing (EGRET), version 0.26.6. All p ≤.3 variables in univariate analyses were included.

RESULTS

From February 1997 to January 1998, 499 patients from the Santos AIDS Reference Center were included in the survey. The studied population consisted of 55.7% men and 44.3% women and had a mean age of 35.8 years. Among participants, 62.3% had completed 8 years of education (the elementary level in Brazil) and 72.3% were ethnically considered as white. Average monthly income per capita was U.S.$159.10 (standard deviation [SD] = U.S.$189.50). Exposure to HIV was believed to have been due to IVDU in 22.2% of subjects, to high-risk heterosexual practice among men in 18.4%, and among women in 38.3%. Among the latter, 74.4% reported having had sexual intercourse with HIV-infected partners and 56.1% with IVDUs. A significant number of individuals had HIV symptomatic disease (31.8% class B and 43.4% class C, according to the CDC's definition) and 56.3% had exhibited CD4⁺ cell counts below 200/mm³ on laboratory follow-up. Seroprevalences of coinfections were as follows: HbsAg, 28.1%; anti-HBc, 46.7%, anti-HCV, -37.0%, and syphilis, -12.9%.

Overall HTLV prevalence was 6.0% (95% CI, 3.9-8.1) for HTLV-1 and 7.4% (95% CI, 5.1-9.7) for HTLV-2 infections, respectively. Of 499 tested samples, 143 were concordantly positive or discordant for HTLV-1/II on EIA screening and therefore selected for confirmatory testing. Western blots yielded 66 positive (46.1%), 42 indeterminate (29.4%), and 32 negative (24.5%) results. Among 66 confirmed HTLV seropositive individuals, 29 (44.0%) exhibited a serologic profile compatible to HTLV-1 infection, 36 (54.5%) to HTLV-2 infection and one (1.5%) dually seropositive individual was found.

Even though both HTLV-1 and HTLV-2 seropositivities occurred more frequently among individuals whose income was <U.S.$136.00, statistical significance in this association was seen for HTLV-2 only. HTLV-1 seropositivity was shown to be associated with <3 years of formal education (p = .05), whereas HTLV-2 seropositivity was significantly associated with black race (p < .01). As to forms of exposure to retroviral infection, univariate analysis demonstrated both HTLV-1 and HTLV-2 seropositivities to be significantly associated with IVDU (OR, 6.34 and 8.11, respectively). No association could be demonstrated between seropositivity for HTLV-1 or HTLV-2 and sexual practice, in terms of lifetime number of sexual partners, past history of two or more STDs, and condom use. However if we consider blood-borne or STD coinfections, HTLV-1 seropositivity was markedly associated with positive anti-HCV, but not with previous HBV infection or syphilis. Similarly, HTLV-2 seropositivity was markedly associated with positive anti-HCV and anti-HBc seropositivity, but not with a positive EIA test result for syphilis (Table 1).

Table 1

Having mutually adjusted all tested variables on multivariate analysis, HTLV-1 seropositivity proved to be independently and significantly associated with IVDU (adjusted OR, 2.99; 95% CI, 1.09-8.20), with less than 3 years of education (adjusted OR, 4.73; 95% CI, 1.56-14.41) and with anti-HCV seropositivity (adjusted OR, 3.03; 95% CI, 1.02-9.01) (Table 2). Similarly, on multivariate analysis, HTLV-2 seropositivity proved to be independently and significantly associated with IVDU (adjusted OR, 3.22; 95% CI, 1.33-7.84), with anti-HCV seropositivity (adjusted OR, 5.40; 95% CI, 1.86-15.66) and with nonwhite race (adjusted OR, 3.32; 95% CI, 1.58-7.00) (Table 3).

Table 2

Table 3

DISCUSSION

Given that HTLV-1 and HTLV-2 share with HIV similar modes of transmission, occurrence of retroviral coinfections in HIV-infected individuals is expected. Overall HTLV seroprevalences in this study, 6.0% for HTLV-1 and 7.4% for HTLV-2, respectively, should be regarded as markedly high in comparison to those seen in other HIV-infected and non HIV-infected populations ⁽¹⁸⁾. Our results therefore indicate intense retroviral exposure of these individuals in Santos. Among Brazilian blood donors, for instance, seroprevalences from 0.1% to 1.8% have been reported in different geographic areas ^(19,20). However, distribution of HTLV-1 and HTLV-2 seropositivities in our cohort contrasts with results from studies conducted either in other Brazilian regions or in the United States. In a cohort of HIV-infected women in the USA, 10.0% of HTLV-2 but only 0.5% of HTLV-1 coinfected study subjects were acknowledged ⁽²¹⁾. In the northeastern region of Brazil, where HTLV-1 is known to be endemic, of a total HTLV seroprevalence of 22.7%, HTLV-1 accounted for 75% of HTLV-seropositive cases ⁽²²⁾.

In population-based serosurveys in highly HTLV endemic areas, both HTLV-1 and II predominantly infect older women ^(23-25). Vertical transmission, more efficacious sexual transmission from men to women and accumulated lifetime risk may account for this particular age and sex distribution. In the present study, however, univariate analysis showed an excess of HTLV-1- and HTLV-2-seropositive individuals among men, even though below statistical significance. This could have been due to a statistically significant predominance of males among IVDUs in the studied population.

Association between HTLV-1 and HTLV-2 seropositivities and socioeconomic parameters has been depicted in previous studies, but not clearly explained ⁽²⁶⁾. Less information about means of retroviral transmission or less access to prophylactic measures against blood-borne and STDs may account for this fact. In the present study, educational background and lower income were associated with HTLV-1 and HTLV-2 seropositivities. Moreover, <3 years of education was shown to be independently associated with HTLV-1 seropositivity (adjusted OR, 4.73; 95% CI, 1.56-14.41). Alternatively, a higher prevalence of IVDU among less wealthy and educated individuals could be speculated. In fact, IVDU was significantly associated with both lower income and educational background in the studied population.

As for HTLV-2 seropositivity, our study showed independent association with black race (adjusted OR, 3.32; 95% CI, 1.58-7.00). In the Caribbean basin, HTLV-1 infection has been similarly associated with black race and therefore related to African slave trade to the American continent ⁽²⁷⁾. Likewise, African-American and Hispanic IVDUs in the United States had a higher frequency of HTLV-2 infection ⁽²⁵⁾.

HTLV-1 and HTLV-2 seropositivities in our cohort are remarkably associated with IVDU. Seroprevalences among intravenous drug users were: 15.9% for HTLV-1 (adjusted OR, 2.99; 95% CI, 1.09-8.20) and 21.0% for HTLV-2 (adjusted OR, 3.22; 95% CI, 1.33-7.84). No significant association was found between HTLV-1 or HTLV-2 seropositivity and sexual practices, as previously reported ^(10,13). However, in our cohort, this lack of association may have been due to a low statistical power (<60%), because few HTLV seropositive individuals were observed among those exposed to sexual practices included among the variables.

HTLV-1 and HTLV-2 seropositivities in this cohort were not shown to be associated with HIV disease progression. However, a cross-sectional survey should not be regarded as the most appropriate for such an evaluation. Chavance et al. ⁽²⁸⁾ argued that greater HTLV-1/II seropositivity in symptomatic HIV-infected patients could simply indicate longer exposure to retroviral infection.

Association between HTLV-1 and HTLV-2 seropositivities and coinfections may allow speculation about mechanisms of HTLV exposure and dissemination in the population. In this regard, syphilis may be taken as a marker of exposure to STDs, seropositivity for HBV as exposure to either STDs or blood-borne infections, whereas seropositivity for HCV predominantly indicates exposure to infected blood. Thus, in our cohort, the independent association between positive anti-HCV and HTLV-1 (adjusted OR, 3.03; 95% CI, 1.02-9.01) or HTLV-2 (adjusted OR, 5.40; 95% CI, 1.86-15.66) seropositivities may suggest exposure to infected blood. Conversely, serologic evidence of past syphilis and HBV infection was not independently associated with HTLV-1 or HTLV-2 seropositivities, which indicates that HTLV acquisition in HIV-infected individuals from Santos does not seem predominantly associated with particular sexual practices.

The independent association between HTLV-1 or II and HCV seropositivities found in the present study raises additional questions because this association might lack clear biologic plausibility. We speculate that HCV seropositivity represents a more reliable surrogate marker of IVDU. In this case, some subjects who denied IVDU when asked about it, could have had their exposure to infected blood revealed by anti-HCV testing. However HTLV-1-infected cell lines have already been shown to support HCV in vitro replication ⁽²⁹⁾, thus raising the possibility of enhancing disease progression. In addition, HTLV-1 and HCV coinfection has been demonstrated to be more frequent among patients with hepatocarcinoma compared with those with chronic HCV infection ⁽³⁰⁾. Further studies are required to clarify this point.

We conclude that even though HIV-infected individuals from Santos, Brazil, are at high risk for HTLV-1 and HTLV-2 coinfections, no significant difference in prevalence of either viral type exists. In our cohort, HTLV-1 and HTLV-2 seropositivities are associated with lower income and educational background. In these individuals, HTLV-1 and HTLV-2 seropositivities are more likely blood-borne and IVDU is their main predictive risk factor. Identifying those population groups at higher risk of acquiring HTLV infection and the main route of viral transmission in this environment will certainly be useful in the establishment of adequate preventive strategies against the dissemination of HTLV-1 and HTLV-2 infections in Santos.

Acknowledgments:

A. Etzel was financially supported by a grant from CAPES, Brazil.

The authors thank Claudio Galperin and Lorena B. Faro for cooperation in serologic assessment of coinfections and Claudio S. Pannuti for critical review of the manuscript.