Although hepatitis C virus (HCV) is primarily transmitted by blood-to-blood contact, evidence for sexual transmission of HCV among heterosexual couples remains controversial.1 Therefore, we read with great interest the study by Terrault et al. reporting on a very low risk of sexual transmission of HCV among 500 monogamous heterosexual couples of which the index person was known to be HCV infected.2 Terrault et al. estimated the minimum and maximum heterosexual transmission rate of HCV based on couples from which the partner was infected with a concordant HCV genotype. Unfortunately, interpretation of these study results is not straightforward because of several potential biases. First, in 2.4% of the included couples, the partner had a history of injecting drug use (IDU); in these couples, it is impossible to exclude HCV infection of the partner through sharing injecting equipment. Second, the minimum HCV transmission rate was based on three HCV concordant couples with significant evidence for highly related viral strains. However, two out of three partners within these couples had a history of either IDU, snorting drug use, or sharing snorting equipment, which makes transmission through drug use more likely than sexual transmission. Therefore, this bias will have lead to an overestimation of the HCV incidence rate. Third, although this study has a cross-sectional design, the researchers report it as a cohort study without having measured the HCV infection status and determinants of interest at the beginning of the sexual relationship. To calculate the HCV incidence rate, the researchers assumed that the index cases were HCV infected before the start of their sexual relationship. However, the HCV infection of the index cases might have occurred during the current relationship. This could have resulted in an underestimation of the incidence rate because too many person-years of exposure were included. Fourth, the investigators excluded couples who had a sexual relationship shorter than 36 months, without providing any specific reason. In addition, couples who had less than three sex acts in the preceding 6 months were excluded, even though they could have had many sex acts in the preceding years. These choices may have lead to a selected study population, which might result in a biased estimate of the transmission risk. To conclude, the study by Terrault et al. is subject to several forms of bias that may have had a substantial effect on the results. Most important, because drug-use-related transmission of HCV was not conclusively excluded, this study is likely to have overestimated the incidence rate of heterosexual transmission of HCV among HCV-monoinfected individuals. BART P.X. GRADY, M.D. MARIA PRINS, PH.D. MAARTEN SCHIM VAN DER LOEFF, M.D., PH.D. Cluster of Infectious Diseases Public Health Service Amsterdam, The Netherlands