The frequency that multiple different subtypes of hepatitis C virus (HCV) simultaneously infect a given individual is controversial. To address this question, heteroduplex mobility analysis (HMA) of portions of the HCV core and envelope 1 region was optimized for sensitive and specific detection of mixtures of HCV genomes of different genotype or subtype. Using the standard HCV genotyping approach of 5'-untranslated region (UTR) analysis, 28 of 374 (7.5%) chronic hepatitis C research subjects were classified as having either multiple-subtype HCV infections (n = 21) or switching HCV subtypes over time (n = 7), the latter pattern implying viral superinfection. Upon retesting of specimens by HMA, 25 of 28 multiple-subtype results could not be reproduced. All three patients with positive results were injection drug users with potential multiple HCV exposures. To address the hypothesis of tissue sequestration of multiple-subtype HCV infections, liver (n = 22), peripheral blood mononuclear cell (n = 13), perihepatic lymph node (n = 16), and serum (n = 19) specimens from 23 subjects with end-stage hepatitis C were collected and analyzed by the HMA technique. Whereas 5'-UTR results implicated mixed-subtype HCV infections in 2 subjects, HMA testing revealed no evidence of a second HCV subtype in any tissue compartment (0 of 70 compartments [0%]) or within any given subject (0 of 23 subjects [0%]). In summary, a large proportion of mixed-genotype and switching-genotype patterns generated by 5'-UTR analysis were not reproducible using the HMA approach, emphasizing the need for additional study.