Background:  Social network data have been used to link infected individuals who engage in risky behaviors associated with hepatitis C virus (HCV) and HIV transmissions. One goal of the Sexual Transmission and Acquisition of HIV Cooperative Agreement Program (SATH-CAP) project in St. Petersburg is to determine whether the chains of referrals built through respondent-driven sampling can help capture the transmission dynamics of diseases like HIV-1 and HCV. The objective of our study was to augment the sociological, behavioral, and biological data collection by performing molecular epidemiologic analyses to compare network linkages with linkages among the viral genomes themselves.

Methods:  The study involves all subjects recruited by respondent-driven sampling in St. Petersburg who are actively infected with HCV. Viral RNA was extracted from serum and amplified by polymerase chain reaction (PCR) using 5' untranslated and core region primers, and the amplified regions were sequenced. A dendrogram was created using the PHYLIP software package to compare the relatedness of the sequences from different individuals. The molecular linkage pattern is compared to that obtained by respondent-driven sampling recruitment.

Results:  Sequences obtained from 77 individuals revealed 3 main genotypes (3a, 1a, and 1b) circulating in the study population, with a preponderance of genotype 3a (62%); and genotypes 1b and 1a were 21% and 17%, respectively. Of the total, 67 samples belonged to 11 recruitment chains of productive seeds or chains with more than 2 individuals; 4 chains with 6, 4, 2, and 3 individuals (excluding seeds) had a single genotype (3a); 4 chains with 7, 11, 5, and 4 members (excluding seeds) had multiple genotypes with >50% of them belonging to 3a; 3 chains with 13, 4, and 2 members (excluding seeds) contained discordant genotypes in variable amounts.

Conclusions:  Our data suggest that molecular epidemiological tools could provide data to support or refute transmission within social networks that are exploited in assembling respondent-driven sampling study populations. The ability of respondent-driven sampling to capture transmission patterns for prevalent infections appears limited, but the 2 data sets combined could provide a more robust exploration of incident transmissions of infectious diseases like HCV and HIV.