In preventive HIV-1 vaccine trials, efficacy may be influenced by the mode of transmission. Prevention of mucosal HIV-1 infection may require lower levels of neutralizing antibody than prevention of infection through exposure associated with injecting drugs, because of compartmentalization, inoculum size, and lack of rapid access to CD4 T cells in the former. Therefore, vaccine strategies that utilize the generation of neutralizing antibodies may be better suited to preventing mucosal transmission of HIV-1 [1,2]. In Thailand, the majority of HIV-1 infections are thought to result from heterosexual transmission , possibly caused by increased female-to-male transmission . In preparation for a phase III field trial of candidate HIV vaccines, it was useful to approximate the proportion of infections that may have been acquired by injecting drug use (IDU) or other means of parenteral transmission . To this end, measurement of antibodies to hepatitis C virus (HCV) may provide a useful surrogate marker for the route of infection, in addition to behavioral risk assessment, since it is almost exclusively associated with parenteral transmission . Furthermore, little information exists about the prevalence of HIV-1/HCV co-infection in community settings in Thailand. Most available data come from samples of select populations of blood donors, sex workers, and IDU [7–12]. In the setting of adult cohort studies preparatory to HIV vaccine trials, designed to estimate HIV-1 incidence, risk factors, and willingness to participate, the prevalence of HIV-1, HCV, and associated risk factors were measured and analyzed in order to estimate the proportion of HIV-1 infection resulting from IDU or other means of parenteral transmission. Another study objective was to assess the validity of self-reported IDU behavior by comparing it with HCV seroreactivity in this setting.
This cross-sectional analysis comprised a subgroup of men and women participating in vaccine trial preparatory cohort studies; participants were recruited from family planning clinics and primary health centers in Rayong and Chon Buri Provinces on the eastern seaboard of Thailand between 1998 and 2001. There were three separate cohorts. The first consisted of 1002 women, aged 20–45 years attending family planning clinics located in three health centers in communities in Rayong Province. The communities consisted of both rural and urban areas with agriculture, fishing, and some manufacturing as the major industries. Women attending the family planning clinics were provided brochures and a description of the project. Those women interested in volunteering completed an informed consent, received education about HIV/AIDS, had pre- and post-test counseling, completed a behavioral risk questionnaire and were tested for HIV and for hepatitis B virus. The second cohort consisted of another 850 women, aged 20–29 years attending the same family planning clinics. The third cohort consisted of 2500 men (48%) and women (52%), 1823 (74.3%) of whom were less than 30 years of age, recruited from four communities in Chon Buri Province. The communities were similar to those in Rayong Province. Recruitment was done by the distribution of brochures, direct mailings, and public announcements. The study procedures were identical for the two provinces. The risk behavioral questionnaire consisted of 34 multipart questions and was administered as a private, structured interview questionnaire with the volunteers by trained community health center staff. Areas covered in the questionnaire included demography (marital status, age, sex, education, occupation), risk behavior (tattoos, transfusions, sexual experience and practices, drug and alcohol use), and knowledge and attitudes about HIV. This questionnaire has been used in phase I/II vaccine trials in Thailand and was originally adapted from an interview protocol for Thai PAVE/HIVNET studies [13,14]. The prevalence of HIV-1 at enrollment was 3.9% (39 women) in the first cohort, 4.0% (34 women) in the second, and 4.8% (121 men and women) in the third. Hence, from these three cohorts, there were a total of 194 prevalent cases (121 women and 73 men) identified. Complete data were available for 191 of these cases. Controls were selected by grouping HIV-positive volunteers by age, sex, and community of residence. The ‘study numbers’ of the HIV-negative volunteers that met those criteria were generated by computer. These study numbers were assigned randomly and not in any order. Thus, if there were five HIV-positive males, between the ages of 25 and 29 from community A, then the first five study numbers from the list of HIV-negative individuals that met the same criteria were chosen. There was one participant for whom no matching control was found. The institutional review boards of the Royal Thai Army Medical Department and the US Army Medical Command approved these studies. Written informed consent was obtained from all study participants.
Individuals were tested for HIV-1 by screening with HIV particle agglutination (Serodia-HIV, Fujirebio, Japan). Repeatedly reactive results were confirmed by Western blot analysis (New LAV Blot I; Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France). Sera were screened for antibodies to HCV using a third-generation enzyme immunoassay (Abbott HCV EIA 3.0, Abbott Laboratories, North Chicago, Illinois, USA). Repeatedly reactive HCV results were considered indicative of infection with HCV. No confirmatory testing was done because of the high specificity of this assay . Enzyme immunoassays were used to test for antibody to hepatitis B core antigen (HBc; MONOLISA anti-HBc, Sanofi Diagnostics Pasteur) and hepatitis B surface antigen (HBs; MONOLISA Ag HBs, Sanofi Diagnostics Pasteur). Passive hemagglutination was used to detect antibodies to surface antigen (Hebsgencell, International Reagents, Japan). All tests were performed according to manufacturers’ instructions. Volunteers without hepatitis B surface antigen or antibodies to HBs antigen were offered vaccination against hepatitis B.
Univariate descriptive analyses were used to assess means, medians, and frequencies, including tests of normality. Bivariate analyses utilizing contingency tables were used to assess associations between HIV-1, HCV, and potentially confounding variables. Fisher's exact test was used for statistical inference for dichotomous variables and Student's t-test for continuous variables. Since no women reported as IDU (compared with men), multivariable analysis of the association between HIV-1 and HCV was modeled separately for men and women, using logistic regression methods. Tests assessed interactions, multicolinearity between HCV and IDU, and model fit. All analyses were performed using Stata Statistical Software, release 6.0 (Stata Corp., College Station, Texas, USA).
The prevalence of reactive antibody to HCV was 8.3% (6/72) in HIV-uninfected men and 4.2% (5/118) among uninfected women (P = 0.34). Among those with HIV-1 infection, the prevalence was 50.7% (37/73) in men and 3.4% (4/118) among women (P < 0.001). Compared with HIV-1-infected women, HIV-1-infected men reported significantly more IDU (34.2 versus 0%; P < 0.001), symptoms or history of sexually transmitted infections (54.8 versus 22.0%; P < 0.01), and tattoos (50.7 versus 5.1%; P < 0.001). Uninfected men also reported significantly more IDU, sexually transmitted infections, and tattoos (data not shown). Among men who reported ever IDU, 96.4% (27/28) were HCV positive. Conversely, among all men who were HCV positive, 62.8% (27/43) reported IDU (Table 1). There were no differences between women from the two provinces for age, education level, and hepatitis B seroreactivity (data not shown). However, there were differences in the number of women married and in the history of sexually transmitted infections, consistent with the differences in the two cohorts (family planning center versus community).
In bivariate analyses (unadjusted) of male volunteers, HCV and IDU were strongly associated with HIV-1, with odds ratios (OR) of 11.3 [95% confidence intervals (CI), 4.4–29.3] for HCV and 12.0 (95% CI, 3.4–41.9) for IDU. Analysis of differences in HCV prevalence between HIV-1 subtypes B and E was precluded by the small number of subtype B infections (data not shown). A history of symptoms or diagnosis of sexually transmitted infections was associated with HIV among both men and women, and, in women only, with the presence of antibodies to HBc (Table 2). Similarly, HCV was strongly associated with IDU (OR, 170.4; 95% CI, 21.6–1343.2) and modestly associated with tattooing and ear piercing among men. (Table 2) Similarly for men, multivariable models with HCV as the dependent variable, only IDU remained significantly associated with HCV (OR, 192.5; 95% CI, 21.3–1736.4). In multivariable models with HIV as the dependent variable, HCV remained strongly associated with HIV, but with a reduced OR (9.7; 95% CI, 2.8–34.0). IDU was no longer significantly associated with HIV (Table 2). The attributable fraction of HCV and IDU associated with prevalent HIV-1 infection was approximately 30% among men. In the multivariable models limited to women, only a history of sexually transmitted infection and antibodies to HBc were significantly associated with HIV infection (Table 2). Potential confounding by cohort (family planning versus community) was not evident after adjustment or stratification. In addition, there was neither significant interactions nor evidence of multicolinearity between covariates. The Hosmer–Lemeshow statistic indicated good model fit.
In this study population, a negative HCV test predicted the absence of IDU. For both men and women, the positive and negative predictive values for HCV seroreactivity and reported IDU were 52% and 99%, respectively. HCV seropositivity and IDU modestly predicted HIV infection: the positive and negative predictive values, respectively, for HIV infection were 79% and 54% for HCV and 89% and 53% for IDU. Among men only, the positive and negative predictive values, respectively, were 63% and 99% for HCV seroreactivity and reported IDU, 86% and 65% for HCV and HIV, and 89% and 59% for IDU and HIV infection.
As a result of a large, coordinated public health effort, Thailand has experienced a marked decline in the incidence of HIV-1 infection since its peak in the early 1990s [3,16,17]. There have been several epidemics of HIV infection in Thailand, including that among Bangkok IDU, which has evolved from being mostly subtype B to subtype E infections [16,18,19]. Concurrently, there is evidence from northern Thailand suggesting that the proportion of HIV infection associated with IDU as a risk factor has increased, perhaps as a result of increased use of injecting drugs concurrent with a reduction in infections caused by sexual transmission . Results from previous cross-sectional studies of HCV in Thailand have varied depending upon the population studied, with estimates ranging from 2 to 6% among blood donors in epidemic HIV areas [7,8], 9.5% among commercial sex workers , and 90% among IDU [11,12]. The differences in HCV prevalence observed in this study and these previous studies can probably best be accounted for by the degree of selection. In particular, the difference between men and women is clearly evident in this analysis. IDU probably accounted for a significant number of the HIV-1 infections among young men in the two provinces under study, given the prevalence and the strength of the association between reported IDU and HCV seroreactivity. This is in contrast with data indicating that parenteral transmission of HIV occurs infrequently among women in these communities, since the prevalence of HCV was similar among infected and uninfected groups, and no women in the study reported ever injecting drugs. This analysis, then, suggests that HIV transmission probably occurs by three patterns in this study population: parenteral transmission among IDU, primarily men; sexual transmission between injecting men and non-injecting women; and sexual transmission between non-injecting sex partners.
This study has several limitations. First, it is cross-sectional and cannot establish a causal relationship between time of exposure and subsequent infection. Second, there is the potential for systematic bias in the use of behavioral questionnaires (one of the reasons for doing this analysis), but this appears to be small, at least for reporting IDU. Third, this analysis used controls matched by variables that included community to minimize confounding caused by the possibility of clusters of HIV and HCV infections within communities. As a result, the total estimated prevalence of HCV among HIV-seronegative controls may be biased and should be interpreted with some caution. Lastly, there is always the possibility of residual confounding.
These results suggest that HCV serologic testing provides a useful marker of IDU risk behavior for HIV-1 infection in populations where IDU have a high prevalence of HCV infection. The importance of these results as related to HIV-1 vaccine efficacy trials may be in the improved characterization of volunteers with intercurrent (`breakthrough') HIV infection, since mode of transmission may be a key determinant of efficacy . Further, these results suggest that IDU may be a common risk factor for HIV-1 infection among young men in Thailand. If this finding is confirmed, national public health policy should increase its focus on behavioral interventions to reduce the frequency of HIV-1 and HCV transmission by IDU.
We wish to thank the cohort study nurses, Vilaiwan Tungsakul, Khanya Pumratana, Sujitra Santativongchai, for their invaluable assistance. Also, we are indebted to the study volunteers, health center staff, the staff of Sattahip District Hospital and Rayong Provincial Hospital, and the Provincial Health Offices of Rayong and Chon Buri Provinces, Thailand, who made this study possible.
Sponsorship: This study was supported by the US Army Medical Research and Material Command.
1.Mascola JR, Frankel SS, Broliden K. HIV-1 entry at the mucosal surface: role of antibodies in protection. AIDS
(Suppl 3): S167–S174.
2.Mascola JR, Stiegler G, van Cott TC, Katinger H, Carpenter CB, Hanson CE, et al
. Protection of macaques against vaginal transmission of a pathogenic HIV- 1/SIV chimeric virus by passive infusion of neutralizing antibodies. Nat Med
3.UNAIDS/WHO. Thailand: Epidemiological Fact Sheet on HIV/AIDS and Sexually Transmitted Infections: 2000 Update.
UNAIDS website, 2000.
4.Mastro TD, Satten GA, Nopkesorn T, Sangkharomya S, Longini IM, Jr. Probability of female-to-male transmission of HIV-1 in Thailand. Lancet
5.Nitayaphan S, Brown AE. Preventive HIV vaccine development in Thailand. AIDS
6.Lauer GM, Walker BD. Hepatitis C virus infection. N Engl J Med
7.Sawanpanyalert P, Boonmar S, Maeda T, Matsuura Y, Miyamura T. Risk factors for hepatitis C virus infection among blood donors in an HIV-epidemic area in Thailand. J Epidemiol Community Health
8.Songsivilai S, Jinathongthai S, Wongsena W, Tiangpitayakorn C, Dharakul T. High prevalence of hepatitis C infection among blood donors in northeastern Thailand. Am J Trop Med Hyg
9.Luksamijarulkul P, Khemnak P, Pacheun O. Human immunodeficiency virus and hepatitis C virus infections among patients attending sexually transmitted disease clinics, Regional 2, Thailand. Asia Pac J Public Health
10.Luksamijarulkul P, Deangbubpha A. Hepatitis C antibody prevalence and risk factors of some female sex workers in Thailand. SE Asian J Trop Med Public Health
11.Apichartpiyakul C, Apichartpiyakul N, Urwijitaroon Y, Gray J, Natpratan C, Katayama Y, et al
. Seroprevalence and subtype distribution of hepatitis C virus among blood donors and intravenous drug users in northern/northeastern Thailand. Jpn J Infect Dis
12.Suriyanon V, Razak MH, Rungruengthanakit K, Maneekarn N, Thongsawat S, Jittiwutikarn J, et al
. HIV and hepatitis C virus infections among drug users in Northern Thailand. Sixth International Congress on AIDS in Asia and the Pacific.
Melbourne, October 2001.
13.Thapinta D, Jenkins RA, Celentano DD, Nitayaphan S, Buapunth P, Triampon A, et al
. Evaluation of behavioral and social issues among Thai HIV vaccine trial volunteers. J Acquir Immune Defic Syndr Hum Retrovirol
14.Nelson KE, Celentano DD, Suprasert S, Wright N, Eiumtrakul S, Tulvatana S, et al
. Risk factors for HIV infection among young adult men in northern Thailand. JAMA
15.Gretch DR. Diagnostic tests for hepatitis C. Hepatology
16.Ruxrungtham K, Phanuphak P. Update on HIV/AIDS in Thailand. J Med Assoc.Thai
17.Kilmarx PH, Supawitkul S, Wankrairoj M, Uthaivoravit W, Limpakarnjanarat K, Saisorn S, et al
. Explosive spread and effective control of human immunodeficiency virus in northernmost Thailand: the epidemic in Chiang Rai province, 1988–99. AIDS
18.Vanichseni S, Kitayaporn D, Mastro TD, Mock PA, Raktham S, Des J, et al
. Continued high HIV-1 incidence in a vaccine trial preparatory cohort of injection drug users in Bangkok, Thailand. AIDS
19.Kalish ML, Luo CC, Weniger BG, Limpakarnjanarat K, Young N, Ou CY, et al
. Early HIV type 1 strains in Thailand were not responsible for the current epidemic. AIDS Res Hum Retroviruses
20.Nelson KE, Eiumtrakul S, Celentano DD, Beyrer C, Galai N, Kawichai S, et al
. HIV Infection in young men in northern Thailand, 1991–1998: increasing role of injection drug use. J Acquir Immune Defic Syndr