Empirical observations underestimate the proportion of human immunodeficiency virus infections attributable to sexually transmitted diseases in the Mwanza and Rakai sexually transmitted disease treatment trials: Simulation results

Sex Transm Dis. 2006 Sep;33(9):536-44. doi: 10.1097/01.olq.0000204667.11192.71.


Background and objectives: Population attributable fractions (PAF) from observational studies may under- or overestimate the contribution of cofactor sexually transmitted disease (STD) to human immunodeficiency virus (HIV) spread. Empirical PAF estimates from the Mwanza and Rakai trials indicated the proportion of HIV infections attributable to STDs was higher in Mwanza than Rakai.

Goal of this study: Estimate the "true" proportion (PAFM) of HIV infections attributable to STDs in the Mwanza and Rakai STD trial populations and explore how the evaluated interventions prevented HIV infections.

Study design: The STDSIM model was used to simulate the 2 populations at the baseline of the trials (with no STD treatment interventions) and counterfactual scenarios in which STD cofactor effects on HIV spread were removed either at the start of the trials or 2, 10, and 20 years into the HIV epidemics. Similar methods were used to quantify the contribution of the cure of each STD to overall HIV impact in each site.

Results: : In Mwanza, the highest PAFM for the effect of any single STD over the 2 years of the trial was due to chancroid (40%). The PAFM for all curable STD was 65%. In Rakai, herpes simplex virus type 2 (HSV-2) was the most important STD (PAFM = 23%); the PAFM for curable STD was 20%. In both sites, the proportion of new infections due to treatable STD decreased over time. The decrease was greater for Rakai, where a behavioral risk reduction that preceded the trial reduced STD prevalence. In both sites, the importance of HSV-2 increased later in the HIV epidemics and STD increased transmission of HIV more than acquisition of HIV. In the Mwanza trial, treatment of chancroid contributed most to preventing new HIV infections.

Conclusions: PAFs calculated from empirical data underestimated the contribution of STD to HIV spread in the Mwanza and Rakai trial populations because STD effects on HIV transmission (as opposed to acquisition) were not captured in the observationally based studies.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Computer Simulation
  • Empirical Research
  • Female
  • HIV Infections / epidemiology
  • HIV Infections / transmission*
  • Humans
  • Incidence
  • Male
  • Models, Statistical
  • Prevalence
  • Sensitivity and Specificity
  • Sexually Transmitted Diseases / epidemiology
  • Sexually Transmitted Diseases / transmission*
  • Tanzania / epidemiology
  • Uganda / epidemiology