Since the 1990s, blood donors have been scanned for anti-hepatitis C virus (anti-HCV) antibodies, which can be defined by enzyme immunoassay as a screening test. In this population, false-reactive ratios have been high. Recently, some authors have aimed to find a cutoff value for anti-HCV different from those established by test manufacturers to predict HCV infection. In this study, 321 patients, after two repeating tests, had reactive results in s/co < 10 titers on anti-HCV test. The patients were 29.6 % (n=95) in women and 70.4 % (n=226) in men. The patients were classified into three groups by Western blot (WB) results (PS, positive; NG, negative; and ID, indeterminate). The average anti-HCV titer of the whole group was 2.61 +/- 1.96. Anti-HCV titers of subgroups were 2.43 +/- 1.95 in NG, 4.93 +/- 2.53 in PS, and 2.50 +/- 1.65 in ID (p< 0.001). There was a significant difference between NG and PS and between PS and ID subgroups (p< 0.001). There was a positive correlation between WB and anti-HCV titers in all patients (r=0.298, p< 0.001), in women (r=0.282, p< 0.001), and in men (r=0.337, p=0.002). According to receiver operator characteristic curve analysis, the cutoff value of anti-HCV titer to predict hepatitis C infection was > 2.61 s/co, with 74.1 % sensitivity and 71.6 % specificity (area under the curve, 0.820; 95 % confidence interval, 0.753 to 0.887). We suggest that an effective cutoff value for anti-HCV other than that established by the manufacturer cannot be assigned to predict hepatitis C infection for blood donors in low-prevalence areas.