BACKGROUND: Sexually transmitted infections (STI) are common in developing countries. The World Health Organisation (WHO) estimates that in 1995, 333 million new cases of syphilis, gonorrhoea, chlamydial infection and trichomoniasis occurred. Human immunodeficiency virus (HIV) infection is also common in developing countries. UNAIDS estimates that over 90% of the 33 million people infected with HIV by December 1999 live in developing countries (UNAIDS 1999). The STI and HIV epidemics are interdependent. Similar behaviours, such as frequent unprotected intercourse with different partners, place people at high risk of both infections, and there is clear evidence that conventional STIs increase the likelihood of HIV transmission. Several studies have demonstrated a strong association between both ulcerative and non-ulcerative STIs, and HIV infection (Cameron 1989, Laga 1993) and there is biological evidence that the presence of an STI increases shedding of HIV and that STI treatment reduces HIV shedding (Cohen 1997, Robinson 1997). Therefore, STI control may have the potential to contribute substantially to HIV prevention. OBJECTIVES: To determine the impact of population-based STI interventions on the frequency of HIV infection, frequency of STIs and quality of STI management. SEARCH STRATEGY: The following electronic databases were searched for relevant randomised trials or reviews: 1) MEDLINE for the years 1966 to current using the search terms sexually transmitted diseases and human immunodeficiency virus infection 2) The Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effectiveness and the Cochrane Clinical Trials Register in the most recent issue of the Cochrane Library 3) The specialist register of trials maintained by the Cochrane Infectious Diseases Group. 4) EMBASE The abstracts of relevant conferences were searched and reference lists of all review articles and primary studies were scanned. Finally, authors of included trials and other experts in the field were contacted as appropriate. SELECTION CRITERIA: Randomised controlled trials in which the unit of randomisation is either a community or a treatment facility. Studies where individuals are randomised were excluded. DATA COLLECTION AND ANALYSIS: Two reviewers independently applied the inclusion criteria to potential studies with any disagreements resolved by discussion. Trials were examined for completeness of reporting. The methodological quality of each trial was assessed by the same two reviewers with details recorded of randomisation method, blinding, use of intention-to-treat analysis and the number of patients lost to follow-up using standard guidelines of the Cochrane Infectious Diseases Group. MAIN RESULTS: Four trials were included. Frequency of HIV infection: In Rakai, after 3 rounds of treatment of all community members for STIs the rate ratio of incident HIV infection was 0.97 (95%CI 0.81 to 1.16), indicating no effect of the intervention. In Mwanza, the incidence of HIV infection in the intervention groups (strengthened syndromic management of STIs in primary care clinics) was 1.2% compared with 1.9% in the control groups (OR=0.58, 95% CI 0.42-0.70), corresponding to a 38% reduction (95%CI 15% to 55%) in HIV incidence in the intervention group. Frequency of STIs: In both Mwanza and Rakai, there was no significant reduction in gonorrhoea, chlamydia, urethritis, or reported STI symptoms among intervention communities. The prevalence ratio of syphilis between intervention and control groups in Rakai was 0.8 (95%CI 0.71-0.89), of trichmoniasis was 0.59 (0.38-0.91), and of bacterial vaginosis was 0.87 (0.74-1.02). In Mwanza, the prevalence of serologically diagnosed syphilis in the intervention community was 5% compared with 7% in the control community at the end of the trial (adjusted relative risk 0.71 (95%CI 0.54-0.93). Quality of treatment: In Lima, following training of pharmacy assistants in STI syndromic management, symptoms were recognised as being due to an STI in 65% of standardised simulated patients (SSPs) visiting intervention and 60% of SSPs visiting control pharmacies (p=0.35). Medication was offered without referral to a doctor in most cases (83% intervention and 78% control, p=0.61). Of those SSPs offered medication, only 1.4% that visited intervention pharmacies and only 0.7% of those that visited control pharmacies (p=0.57) were offered a recommended regimen. Similarly in only 15% and 16% of SSP visits respectively was any recommended drug offered. However, education and counseling were more likely to be given to SSPs visiting intervention pharmacies (40% vs 27%, p=0.01). No SSPs were given partner cards or condoms. In Hlabisa, following the intervention targeting primary care clinic nurses (strengthened STI syndromic management and provision of STI syndrome packets containing recommended drugs, condom, partner cards and patient information leaflets), SSPs were more likely to be given recommended drugs in intervention clinics (83% vs 12%, p <0.005) and more likely to be correctly case managed [given correct drugs, partner cards and condoms] (88% vs 50%, p <0.005). There were no significant difference in the proportions adequately counseled (68% vs 46%, p=0.06), experiencing good staff attitude (84% vs 58%, p=0.07), and being consulted in privacy (92% vs 86%, p=0.4). There was no strong evidence of any impact on treatment seeking behaviour, utilisation of services, or sexual behaviour in any of the four trials. REVIEWERS' CONCLUSIONS: There is limited evidence from randomised controlled trials for STI control as an effective HIV prevention strategy. Improved STI treatment services have been shown to reduce HIV incidence in an environment characterised by an emerging HIV epidemic (low and slowly rising prevalence), where STI treatment services are poor and where STIs are highly prevalent. There is no evidence for substantial benefit from treatment of all community members. There are however other compelling reasons why STI treatment services should be strengthened and the available evidence suggests that when an intervention is accepted it can substantially improve quality of services provided. Further community based randomised controlled trials that test a range of alternative STI control strategies are needed in a variety of different settings. Such trials should aim to measure a range of factors that include health seeking behaviour and quality of treatment as well as HIV, STI and other biological endpoints.