Ng BE, Butler LM, Horvath T, Rutherford GW. Population-based biomedical sexually transmitted infection control interventions for reducing HIV infection. Cochrane Database Syst Rev 2011 March 16;3.
Transmission of sexually transmitted infections (STIs) and HIV infection are related, as similar behaviors (such as frequent unprotected intercourse with different partners) lead to high risk of STI infection and HIV infection. Certain STIs also increase the likelihood of HIV transmission. STI interventions applied to populations at high coverage levels may have a mass effect on reduction in the level of STI transmission in the population, and contribute substantially to HIV prevention by reducing HIV shedding among HIV-positive persons with an STI, and reducing susceptibility of HIV-uninfected persons with an STI to acquire HIV.
The objective of this Cochrane systematic review was to determine the impact of population-based biomedical STI interventions on the incidence of HIV infection in the general population. Published studies and conference abstracts of randomized controlled trials in which the unit of randomization was either a community or a treatment facility and in which the primary outcome was incident HIV infection were included. Meta-analysis was performed.
PubMed, EMBASE, CENTRAL, Web of Science/Social Science, PsycINFO, LILACS were searched for reports published between January 1, 1980-August 16, 2010. The AEGIS database of conference abstracts from CROI, IAC, and IAS meetings was searched from conference inception dates through 2007, with manual searches of the conference web sites performed for abstracts presented through 2010. Both www.clinicaltrails.gov at NIH and the Meta-Register of Controlled Trials (www.controlled-trials.com/mrct) were searched for ongoing research. Reference lists of all studies were reviewed.
This is an update of an existing Cochrane review from 2001,(1) originally updated in 2004.(2) The search methods were updated, and the inclusion/exclusion criteria were modified to focus on biomedical interventions and on HIV incidence as a primary outcome. Behavioral interventions, eligible for inclusion in the reviews previous versions, were excluded from this updated version. One new study was included since the 2004 update, and two studies were excluded that were in the 2004 review.
Eligible for inclusion were randomized controlled trials in which the unit of randomization was either a community or a treatment facility; the intervention was a population-based biomedical STI program, including improved STI management and mass antibiotic treatment of populations; and in which the primary outcome was HIV incidence.Community was interpreted to include a group of villages, an arbitrary geographical division, or the catchment population of a group of health faciltities. Among studies meeting exclusion criteria were trials in which individuals, and not communities, were randomized; trials based in sub-populations at risk such as sex workers, injection drug users, or men who have sex with men, and trials with interventions that were primarily behavioral.
The unit of analysis was individuals. The unit of randomization in each study was communities. Data from cluster-randomized trials included in this review included calculation of the intracluster correlation coefficient (ICC) for each study. The ICC is used to calculate the design-effect, which is then used to calcuate the effective sample sizes of intervention and control groups.(3) The ICCs for the primary outcome were found for 3 of 4 studies. For the remaining study, it was approximated by using the ICC for one of the other studies, as the design and intervention were most similar for these two. Where interventions and study populations were similar across studies, outcomes were pooled and examined using a fixed-effects model when there was low heterogeneity, and a random-effects models where there was moderate or high heterogeneity between studies.
General adult populations and people with STIs.
Any population-based biomedical STI intervention including, but not exclusive to 1) improved STI treatment services and 2) mass treatment of whole communities (i.e., all individuals, both symptomatic and asymptomatic, in a community, who met the study's criteria for treatment of STI).
Primary outcome was incidence of HIV infection in a representative community sample. Secondary outcomes included 1) incidence or prevalence of STIs in a representative community sample, and 2) safer sexual behaviors in a representative sample, including a) proportion of people reporting a reduction in rate of partner change, b) regular condom use with casual partners and c) two or more sexual partners in the past year.
The search initially identified 6,003 results, of which 776 were duplicates, and an additional 3,268 were excluded by one author as clearly irrelevant. Of the remaining 1,959 citations and abstracts, 46 articles were chosen for full text review, of which four studies were included in the review. Two reviewers independently abstracted data from eligible studies.
Four trials were included in the analysis. One trial evaluated intensive STI control via home-based mass antibiotic treatment in Rakai, Uganda. The other three trials evaluated combinations of improved syndromic STI case management in clinics, STI counseling, and STI treatment in Tanzania, Uganda, and Zimbabwe.
Meta-analysis of 23,981 participants in the four studies showed no evidence that biomedical STI interventions reduced the incidence of HIV (RR=0.93 [95%CI 0.67-1.3]), with moderate heterogeneity between the studies. Combining the three studies with improved STI management (not mass treatment), meta-analysis of 18,194 participants found no evidence that these interventions reduced the incidence of HIV (RR=0.88, 95%CI=0.52-1.5) with high heterogeneity between studies. The study evaluating mass STI treatment among 5,787 participants found no evidence of effect of this intervention on HIV incidence. (RR=1.02 [95%CI 0.71-1.46])
Data were combined from the two studies that evaluated syphilis, gonorrhea and chlamydia prevalence. There was a significant 12% reduction in the prevalence of syphilis among those receiving the biomedical intervention compared to the control group (RR=0.88, [95% CI 0.80-0.96]) with low heterogenetiy between the studies, and a statistically significant 51% reduction in prevalence of gonorrhoea (RR=0.49 [95%CI 0.31-0.77]), with low heterogeneity between the two studies. For chlamydia, there was no significant difference between the biomedical intervention and control group (RR 1.03 [95% CI 0.77-1.36]), with low heterogeneity between studies.
Data were combined for three studies with a significant 18% increase in regular condom use with casual partners in those exposed to any biomedical intervention (RR 1.18 [95% CI 1.04-1.33]) compared to control, with high heterogeneity among the studies. Data from all four studies found a small, but significant 7% increase in the percentage of people who had two or more sexual partners in the past year (RR 1.1 [95%CI 1.01-1.1]), with low heterogeneity among the studies.
Evidence from community randomized controlled trials suggests that biomedical STI control interventions are not an efficacious HIV prevention strategy. There is no evidence of efficacy in populations where HIV incidence is not rising rapidly. Improved STI treatment services were shown in one study to reduce HIV incidence in an environment characterized by an emerging HIV epidemic (low and slowly rising prevalence), where STI treatment services were poor and where STIs were highly prevalent. Meta-analysis of these four studies found no statically significant effect on HIV incidence. However, strengthening the quality of STI services can lead to decreased prevalence of some STIs.
All relevant PRISMA criteria were met.
STI interventions have minimal population effect on HIV incidence in populations where HIV has become endemic. Strengthening of effective STI services can improve quality and lead to decreased prevalence of some STIs.
- Wilkinson D, Rutherford G. Population-based interventions for reducing sexually transmitted infections, including HIV infection. Cochrane Database of Systematic Reviews 2001.
- Sangani P, Rutherford G, Kennedy GE, Wilkinson D. Population-based interventions for reducing sexually transmitted infections, including HIV infection. Cochrane Database of Systematic Reviews 2004.
- Kerry SM, Bland JM. The intracluster correlation coefficient in cluster randomisation. BMJ.1998 May 9;316(7142):1455 Review.