HIV-1 epidemic in Russia: an evolutionary epidemiology analysis

Abstract

BackgroundThe emergence of HIV-1 in Russia came soon after the collapse of the Soviet Union. With more than half a million cases within 5 years, the post-soviet Russian epidemic is one of the fastest and largest HIV outbreaks ever recorded. Although driven by intravenous drug users (IDUs), why the epidemic spread so rapidly and on a large scale is unclear. One popular practice between 1990 and 2000 was to use blood as a buffer or as a purifier of homemade narcotics; this practice could account for the massive spread of HIV among Russian IDUs. We aimed to assess whether the epidemic was driven by contamination of narcotics at their source rather than by needle sharing. MethodsWe have previously shown that by comparing number of prevalent cases and skyline plot we can recover epidemiologically coherent combinations of transmitters and generation time. Briefly, the ratio of the number of prevalent cases divided by the skyline plot estimate (product of the effective number of infections with the generation time) is equal to the ratio of the variance of secondary infections divided by the generation time. We assume that the infected population is divided into non-transmitters and transmitters with Poisson-distributed secondary infections. We investigated the Russian epidemic using molecular sequences and number of prevalent cases that we had systematically collected from previous surveys, and analysed the variation of transmissibility. FindingsWe found that the scenario of point transmission is highly unlikely: if a low proportion of transmitters (<5%) was operating throughout the epidemic, they could have a large number of secondary HIV infections (>160 per infected person) but only throughout a large time period (>6 months), rendering the epidemic a longitudinal rather than an episodic event. The most epidemiologically coherent scenario is that the generation time was contracted down to about 1 month with 30–60% transmitters and 8–27 secondary infections on average per transmitter. This finding means that most of the IDUs had transmitted HIV many times through their lifetime, at least half of which was during their primary infection. InterpretationIn Russia, because of the high viral load of acute infection, the high number of contacts per IDU led to a boosting of expansion to a moderate-to-high number of transmissions (four to 14) during the first month of infection. In the absence of any prevention, there was nothing to decelerate the dispersal of blood-borne pathogens. As soon as HIV was introduced, the outbreak expanded uncontrollably. We show that such an event does not require a massive amount of centrally distributed, contaminated product, but a large number of strongly connected transmitters. FundingUK Medical Research Council, European Commission, National Institute on Drug Abuse grant DP1 DA034989 (Preventing HIV Transmission by Recently-Infected Drug Users).