Within the framework of hepatitis C virus (HCV) prevalence monitoring, we evaluated oral fluid (OF), which is richer in IgG than whole saliva, as a possible alternative to serum for the detection of HCV antibodies. Paired OF and serum samples were collected from 90 individuals, including 45 HCV-positives and 45 HCV-negatives. The detection of HCV antibodies in both serum and OF was performed using the Ortho HCV 3.0 SAVe enzyme-linked immunosorbent assay (ELISA) (Ortho-Clinical Diagnostics, Inc., Raritan, NJ), but a modified, more sensitive protocol was used to process OF. The sensitivity and specificity of this assay were 86.67% (95% confidence interval (CI): 72.51-94.46%) and 100% (95% CI: 90.20-99.80%) in OF and 100% in serum. The correlation obtained between both types of clinical specimens was excellent (k: 0.87, 95% CI: 0.66-1.07). However, the negative predictive value (NPV) of the assay in OF decreased with the prevalence of HCV infection in the population studied. Our results suggest that the modified Ortho HCV 3.0 SAVe ELISA is suitable for the detection of HCV antibodies in OF for epidemiological studies. Using this assay, we observed an unadjusted anti-HCV prevalence of 78.6% among a population of intravenous drug users; when adjusted to account for assay sensitivity, this prevalence may be closer to 90%.