The HIV epidemic in Thailand has evolved through several stages over time. Commercial sex was the main route of transmission during the epidemic in the early 1990s, which was successfully controlled by the government’s 100% condom programme.1 However, a recent modelling study estimated that the general low-risk heterosexual population accounted for 43% of all new infections in 2005.2 Prevention strategies must, therefore, be adapted to engage people with HIV in reducing further HIV transmission.

Public health experts widely emphasise the importance of addressing HIV prevention with people with HIV while scaling up access to antiretroviral (ARV) treatment.3 4 These prevention approaches, often termed positive prevention, include screening and treatment for sexually transmitted infections (STIs), distribution of free condoms, risk-reduction counselling, assistance to people with HIV in notifying partners of their HIV infection and partner HIV testing.3 As positive prevention has only recently become an international priority in HIV prevention, few strategies to reduce sexual risk behaviour among people with HIV have been evaluated for effectiveness in randomised controlled trials.5 6 Implementation has been limited to industrialised countries5 6 or to settings with temporary additional resources.7 The Thai Ministry of Public Health (MOPH) provides universal access to ARV treatment under the Royal Thai Government national healthcare programme.8 Resources to incorporate prevention services into HIV care and treatment are limited, so factors should be identified to design a targeted intervention.

An earlier report by Sirivongrangson et al showed frequent STIs and sexual risk behaviour among Thai women with HIV, especially among those attending STI clinics.9 However, the need for positive prevention services among people with HIV, including men, attending HIV care remains to be assessed. We assessed the risk for HIV transmission among people with HIV attending the National Infectious Disease Institute in Bangkok, Thailand. The prevalence of several STIs and sexual risk behaviours were assessed to explore ways to target STI screening and behavioural interventions.

METHODS

Project design

Baseline data from people with HIV seeking care under the national HIV treatment programme8 at the National Infectious Disease Institute of the Thai MOPH (Bamrasnaradura Institute, Nonthaburi) were assessed. Men entered the project at the Department of Medicine and women entered the project at the Department of Gynecology. Consecutive patients were included in the assessment as time and staff availability allowed; all assessments were done by two designated staff in each clinic. The project ran from July 2005 to September 2006.

Demographic and behavioural information was collected in a standardised way as part of routine care at a single time point for each patient assessed. Data about age, sexual partners, male-to-male sexual behaviour and condom use were collected during face-to-face interviews with a counsellor or nurse. People were asked about their types of partnerships in the last 3 months; a steady partnership was defined as a sexual relationship lasting at least 2 months with an emotional bond between partners, a casual partnership was without such a bond and a commercial partnership was defined as a sexual relationship with the exchange of money or goods. The presence of STI-related symptoms was recorded, including vaginal or urethral discharge, genital pain or lesions, dysuria for men and lower abdominal pain for women.

All women underwent a pelvic examination and endocervical specimens were collected for STI testing. Men were only examined if they reported STI-related symptoms; all men provided urine samples for STI testing. STI-related examination findings for women were defined as genital ulcer, inflamed cervix, yellow or green vaginal discharge and adnexal or cervical motion tenderness. STI-related examination findings for men were genital ulcer, urethral discharge and swollen testis. STI treatment was conducted according to the guidelines of the STI division of the Thai MOPH and was provided free of charge along with condoms. Risk reduction and HIV disclosure counselling were provided as part of routine HIV care. Partners of people with HIV were offered a HIV test at the same clinic, CD4 cell count test and access to HIV care as appropriate; testing was provided free of charge if the partner was not able or willing to pay. People with HIV received opportunistic infection prophylaxis and ARV treatment according to the Thai national guidelines for HIV care (ARV provided to patients with CD4 cell count <200 cells/mm3 or symptoms of HIV infection).10 People were asked about ARV treatment adherence by recall of doses taken in the last 3 days. Condoms were available at the clinics and provided upon request by patients.

Laboratory evaluation

First-void urine was collected from men and endocervical specimens from women for the detection of Chlamydia trachomatis and Neisseria gonorrhoeae by Amplicor PCR analysis (Roche Diagnostic System, Basel, Switzerland). Serum samples were tested for syphilis using the rapid plasma reagin (RPR) card test (New Market Laboratory, Kentford, UK) for screening and the Treponema pallidum haemagglutination assay (TPHA; Fuji-rebio Inc, Tokyo, Japan) for confirmation. Vaginal swab specimens were collected from women for saline wet preparations for the detection of T vaginalis by light microscopy.

HIV infection status was confirmed by Murex HIV1-2 ELISA (Murex Biotech Ltd, Dartford, UK) if not previously documented. CD4 cell count testing was performed using a FACScan flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, California, USA).

Data analysis

Demographic, behavioural and clinical characteristics of men and women were compared. Unsafe sex was defined as unprotected last sex with a partner with negative or unknown HIV status. People were considered to have a STI if they had clinician-confirmed genital ulcers or positive test results indicating chlamydial infection, gonorrhoea, trichomoniasis or reactive syphilis serology with confirmation by TPHA. Factors associated with unsafe sex, STIs and HIV disclosure were analysed in bivariate and multivariate models; continuous variables were dichotomised at median levels.

Data were analysed and statistical tests performed using SPSS (v.12). The χ2 tests were used to test for differences in proportions. Frequency data were analysed by χ2 test or χ2 test for trend. Logistic regression was used to calculate odds ratios (OR) with 95% CIs. Multivariate modelling was carried out using multiple logistic regression with the final choice of model based on predictive value as measured by the index area under the receiver-operator curve; risk factors from bivariate analysis with p<0.2 were included in the initial models. Factors that were highly correlated with other factors in the model were examined to identify the best factor to include in the final multivariate model.

RESULTS

From July 2005 to September 2006, data about HIV transmission risk were assessed among 894 people with HIV (395 men and 499 women) seeking routine HIV care. Participants represented 12.4% of the 3188 men and 50.0% of the 999 women who attended the hospital for HIV care during that time period. Their median age was 36 years (range 19–74 years). Women had known their HIV status for twice as long as men (table 1).

Only 72.1% (178/247) of the sexually active men were asked about male-to-male sexual behaviour; of these men, 24.7% disclosed having had sex with another man in the last 3 months (table 1). Among those with steady partners, 45.1% of the men and 8.3% of the women also had casual or commercial partners. Around half of the steady partners accepted an HIV test; 45.5% of female partners of men with HIV and 33.3% of male partners of women with HIV tested HIV negative. Of the 138 people with steady partners who tested HIV positive, 78.6% reported condom use during last sex. More sexually active women (55.3%; 130/235) requested condoms than did men (14.2%; 29/204; p<0.001).

More women were receiving ARV treatment (75.1% vs 50.7%; p<0.001), had received treatment for a longer duration and had higher CD4 cell counts (table 1).

Factors associated with unsafe sex

Overall, 364 (40.7%) people with HIV reported that they had one or more partners with negative or unknown HIV status: these partners were 263 steady partners, 127 casual partners, 43 sex workers and 6 clients of sex workers. Unsafe sex (unprotected last sex with partner of negative or unknown HIV status) was more common with casual, commercial and male-to-male sex partners compared with steady heterosexual partners; people on ARV treatment were less likely to engage in unsafe sex. Both factors remained statistically significant in multivariate analysis (adjusted odds ratio (AOR) 2.1; 95% CI 1.1 to 4.1 and AOR 0.3; 95% CI 0.1 to 0.6, respectively; table 2). The variable HIV disclosure was dropped in the multivariate model as this was highly correlated with ARV treatment (p = 0.02).

Factors associated with STI detection

STI prevalence was non-significantly higher among people <25 years of age compared with those >25 years of age (12.1% vs 8.6%; p = 0.5), with only 33 people in the younger age group. People on ARV treatment were less likely to have a STI compared with people not on ARV treatment (7.2% vs 11.5%; p = 0.03) (table 3). Adjusting for gender in the multivariate analysis, ARV treatment was no longer significantly protective for STI infection (AOR 0.7; 95% CI 0.4 to 1.1).

Factors associated with HIV disclosure to steady partners

In multivariate analysis, HIV disclosure to steady partners was more common among women than men (82.5% vs 65.9%; AOR 2.0; 95% CI 1.0 to 4.0) and among people on ARV compared with those not on ARV treatment (80.8% vs 66.7%; AOR 2.7; 95% CI 1.4 to 5.3). Also, those reporting protected sex with their steady partners were more likely to have disclosed their HIV status (82.7% vs 62.7%; AOR 2.5; 95% CI 1.2 to 5.2) (table 4).

DISCUSSION

This report shows substantial risk for sexual HIV transmission among people attending HIV care in Thailand. Sexual risk behaviour was common and a significant proportion of people with HIV had not disclosed their HIV status to their steady partners. This information is timely as the Thai government is scaling up ARV treatment to reach universal coverage. The number of people with HIV on ARV treatment has increased sharply since 2003,8 leading to a large population of people living with HIV/AIDS in Thailand who could potentially transmit HIV to others.

Sexual risk behaviour was common among people attending HIV care. Men with HIV frequently had unprotected sex in casual and commercial partnerships and only 14% of men requested condoms at the clinic. In addition, nearly 25% of men who were asked reported sex with men in the last 3 months. Since a quarter of men were not asked about same sex behaviour, the true prevalence of this behaviour among patients screened is unknown. Addressing positive prevention among men who have sex with men (MSM) is critical as the HIV prevalence among MSM in Thailand is increasing rapidly.11 Unprotected sex with a partner of negative or unknown HIV status was more common with casual, commercial or male-to-male sex partners than with steady heterosexual partners. Positive prevention efforts should aim to reduce transmission to all partner types and reduce the number of partners.

ARV adherence and HIV transmission risk are important to monitor during ARV treatment scale-up. We found high levels of self-reported ARV adherence, consistent with a previous report among 149 persons attending ARV care at the same HIV-care institute.12 A smaller proportion of people on ARV treatment in our project reported unsafe sex compared with those not on treatment, echoing findings from US studies,13 although some studies have reported increased sexual risk behaviour among people on ARV treatment.14 15 A cohort study in Uganda found a decrease in unsafe sexual behaviour among people with HIV infection after starting ARV treatment.7 People on ARV treatment in our project could reflect a population with overall healthier behaviour, including sexual and healthcare seeking behaviour. Also, patients on ARV treatment have more frequent clinic visits than those not on ARV and so are exposed to counselling messages that might have resulted in increased condom use. Risk behaviour among people starting on ARV treatment should be monitored longitudinally to evaluate the impact of treatment on continued risk for HIV transmission.

The low gonorrhoea and chlamydia prevalence makes universal screening for people with HIV impractical in a setting with limited resources. Targeted gonorrhoea and chlamydia screening for people with high-risk behaviour should be considered, along with periodic cross-sectional assessments to monitor the STI prevalence. Syphilis prevalence was somewhat higher, especially among men. Given the low cost of syphilis screening and the potential for serious complications of untreated infection, all patients with HIV should be screened. The prevalence of STIs appears to be substantially higher among people with HIV younger than 25 years old compared with those older. The younger group comprised <5% of our sample, limiting our power to detect a significant difference and to analyse factors associated with STIs in this group; further evaluation of STI screening needs in this population is warranted.

Fewer men than women disclosed their HIV status to their steady partners. However, only half of the men with a steady partner were asked about disclosure, suggesting that disclosure counselling for men needs to be strengthened. People receiving ARV treatment were more likely to disclose HIV to their steady partners and people who had disclosed were less likely to report unprotected sex with their steady partner.

The HIV discordance rate was high among steady partners accepting an HIV test indicating that significant prevention opportunities exist. Because around half of all new HIV infections in Thailand are estimated to occur among partners of people with HIV,2 discordant couples need to be identified and targeted for prevention efforts.

Data were collected as part of routine clinical practice; therefore, patients included the broad population of people attending HIV care rather than being a selected study population. However, this approach has some limitations. First, data were only collected when staff were available to complete the clinic form and to collect STI samples, leading to modest coverage of the hospital’s patient population and creating the possibility of selection bias towards people most willing or able to participate in the project. Men and women entered the project through different clinics, so it is possible that this contributed to differences observed between men and women, including the substantial difference in proportion receiving ARV treatment. Second, some clinic forms were incomplete possibly due to staff time constraints, staff discomfort asking some questions or unknown reasons. This may have led to biased frequency results for male-to-male sexual behaviour and HIV disclosure among men as mentioned earlier, although risk factor analysis for those with complete data remains valid. The incompleteness of data, particularly on sensitive topics, suggests that these topics were not always discussed; this provides useful direction for strengthening counselling services. Finally, we did not address HIV transmission risk through injecting drug use. Prevention measures for this risk behaviour should be included in prevention services for people with HIV.

Our findings indicate a need to reduce HIV transmission risk among persons attending HIV care, especially among those not on ARV treatment and among men with casual, commercial and same-sex partners. These prevention services need to be addressed within the limits of the Thai public healthcare system, which already faces high demands because of the universal access ARV treatment programme. Creative and sustainable behaviour change communication interventions and materials need to be designed to reduce HIV transmission risk behaviours among people with HIV. Brief risk-reduction messages provided by clinicians at every visit have been proven effective16 and may be considered for routine HIV care in Thailand. The Thai MOPH is working towards integration of positive prevention into routine HIV care, including provision of free condoms for all sexually active people with HIV, and will evaluate the effectiveness of the programme in reducing HIV transmission risk.