Paper # 997 
HAART Eliminates HIV Epidemics in a Network Model of an IDU Community
P Bastani1, E Druyts2, R Hogg2, B Marshall2, J Montaner2, B Ramadanovic1, A Rutherford1, Krisztina Vasarhelyi*1, and IMPACT-HIV
1IRMACS Ctr, Simon Fraser Univ, Burnaby, Canada and 2BC Ctr for Excellence in HIV/AIDS, Vancouver, Canada
Background: HIV spreads rapidly among injecting drug
users (IDU) who share contaminated syringes. However, early epidemics can be
reversed, and HIV prevalence decreases substantially in established epidemics,
when interventions are broadly implemented. In a cellular automaton model, we
previously showed that HIV is eliminated in IDU groups, if transmission due to
risk behavior is reduced below a threshold. Here, we evaluate Highly Active
Antiretroviral Therapy (HAART) as an intervention to reduce HIV incidence and
prevalence in IDU communities.
Methods: We use a network model based on a
Mover-Stayer framework and our previous cellular automaton model to evaluate
HAART as prevention. In the model, IDU are distinguished based on
syringe-sharing behavior and HIV status, and exert social influence on peers,
encouraging, or discouraging syringe sharing. We applied HAART at coverage
levels of 0% to 100%, assuming complete adherence and no drug resistance,
tracked HIV incidence, and prevalence to equilibrium. Community composition,
needle sharing frequency (60/month), and initial HIV prevalence (31%) were
derived from data on IDU enrolled in the Vancouver Injection Drug User Study
(VIDUS). Published transmission rates for HIV disease stages were used. HAART,
initiated after 5 years (Scenario 1), was combined with reduced risk behavior
(Scenario 2), the latter repeated with HAART initiated after 1 year (Scenario
3).
Results: Without intervention, HIV spreads rapidly
and reaches very high prevalence (90%) in the model. With increasing HAART
coverage, HIV incidence and prevalence decrease for all scenarios, eventually
reaching 0%. Without change in risk behavior (Scenario 1), HIV prevalence
decreased gradually to 60% HAART coverage, dropping rapidly thereafter. Behavioral
interventions (Scenarios 2 & 3) amplified HAART effects. At 40% to 50%
HAART, both incidence and prevalence were reduced by about half. Above 80%
coverage, the epidemic was effectively eliminated. Early HAART initiation
showed little impact.
Conclusions: High prevalence at low HAART coverage
confirms that HIV spreads efficiently on IDU networks. Simulation results
suggest that HAART, especially combined with behavioral interventions, is
highly efficient in controlling the epidemic, and can potentially eliminate it.
Treatment variation, early HAART initiation in the acute phase, behavioral and
demographic complexities and features of specific epidemics will be considered
in future studies.


|