Article Text


Epidemiology oral session 11: Various topics of special interest
O1-S11.01 Time evolution of the fraction of new HIV infections due to primary infection among high risk groups in southern India
  1. M Pickles1,
  2. M C Boily1,
  3. P Vickerman2,
  4. B M Ramesh3,
  5. R Washington4,
  6. K Deering5,
  7. S Verma3,
  8. J Blanchard6,
  9. S Moses6,
  10. M Alary7
  1. 1Imperial College London, UK
  2. 2LSHTM, UK
  3. 3Karnataka Health Promotion Trust, Bangalore, India
  4. 4St John's Research Institute, Bangalore India
  5. 5University of British Columbia, Canada
  6. 6University of Manitoba, Winnipeg, Canada
  7. 7Centre hospitalier affilié universitaire de Québec, Quebec, Canada


Background HIV infectivity is considerably higher during primary infection (PI: first few months after infection). PI is believed to be an important biological driver of HIV transmission at the population level, which has implication for prevention. We aimed to estimate the contribution of PI to HIV spread among high-risk groups in southern India.

Methods We used incidence predictions from a bespoke dynamical model of HIV transmission among FSW/clients parameterised and fitted to district-specific behavioural data and HIV/STI prevalence estimates, within a Bayesian framework. Multiple rounds of cross-sectional survey data from among FSW/clients of Mysore and Belgaum, carried out as part of the evaluation of Avahan, the India AIDS initiative, were used to inform plausible parameter ranges and estimate HIV prevalence. Assumed risk of HIV infection (RRHIV/PI) during PI was elevated by 4.5- to 19-fold compared to asymptomatic infections (based on systematic review) and lasted for 3 to 6 months. The annual fraction of new HIV infections (AF) due to primary infection was estimated over time from the 196 and 796 posterior parameter sets fitting Mysore and Belgaum data, respectively.

Results The median AF (overall: clients+FSW) declined from 57.3% (min=34, max=82)% in 1987 (assumed start of epidemic) to 49% (29,70)% in 2000, 43.7% (25,65)% in 2004, 23.8%/24.3% (10/10, 42/43)% in 2010/2025 for Mysore (Abstract O1-S11.01 figure 1). Similar AF estimates were obtained in Belgaum and for FSWs and clients separately in both districts. The univariate association between AF and RRHIV/PI declined slightly over time (Abstract O1-S11.01 figure 1). In multivariate regression analysis, RRHIV/PI (slope=0.027), duration of the whole infectiousness period (slope=−0.014) and duration of primary infection (slope=0.587) were significantly associated with AF early in the epidemic (1988) (R2=0.833, p value<0.001). However, the AF in 2025 was also associated with number of years selling sex by street-based FSW (slope=−0.020, p value=0.042) and number of years buying sex by high-activity clients (slope=−0.004, p value=0.055) (R2=0.684).

Abstract O1-S11.01 Figure 1

FSW and clients combined Mysore.

Conclusions The role of primary infection decreased as the HIV epidemic matured but could still account for a large fraction of new infections, especially if RRHIV/PI is above 10. Early on, its contribution depended on parameters of primary infections. Later on, its contribution also depended on the renewal of high-risk susceptible population, which fuels HIV incidence.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.