. Cook et al. did not find that whole blood was a more sensitive source of hepatitis C virus (HCV) RNA than plasma. Careful review of their paper revealed several potential reasons for this discrepancy, including the following. (i) A different method of extracting RNA from whole blood was used (3). We did not use Trizol following extraction with Catrimox (4, 9). The pH of the phenol-guanidinium mixture in Trizol is not published. We found that the phenol pH is critical for optimal extraction (4, Brashear et al., 10th Triennial Int. Symp. Viral Hepatitis Liver Dis. 4). (iv) Finally, the serum volume was not adjusted to the whole-blood volume (3), which we took into account (6).As suggested by Cook et al., intracellular HCV RNA is a potential explanation for the improved sensitivity of HCV RNA detection in whole blood (3), yet we found that a significant portion of HCV RNA in the crude cellular pellet was not intracellular (6). Our study of 115 patients with HCV infection demonstrated that the improved sensitivity of whole-blood HCV RNA detection correlated with the presence and quantity of cryoglobulins (8). We also noted highly significant differences in the detection of HCV RNA for whole blood and plasma among 52 interferon-treated patients who were monitored throughout treatment (7). As was the case for the one patient studied by Cook et al. (3), among our 52 patients the clearance of intracellular virus occurred nearly simultaneously with the clearance of plasma virus during therapy (7).The extraneous bands described by the authors (3) are not routinely present in our work (e.g., see Fig. 1, 3, and 5B in reference 4). In addition, Cook et al. stated that it was not tenable that our antibody-negative subjects were truly infected with HCV since we did not evaluate liver biopsy samples for HCV RNA (3). However, we proved that several HCV antibody-negative patients had HCV RNA present in multiple whole-blood and plasma samples collected over months and years, using primers from several different regions of the genome (4, 6). In addition, we confirmed our results by Southern blotting and sequence analysis of PCR products and by direct comparison with results obtained with the use of commercial HCV RNA assays (4, 6, 9). The finding of prolonged RNA positivity in the absence of detectable HCV antibody has been described by others (2), including after the experimental infection of chimpanzees (1). In conclusion, we believe that the difference between our work and that of Cook et al. (3) relates to differences in the whole-blood RNA extraction methodology. Authors' Reply Drs. Schmidt and Stapleton think that the difference between our work (3) and theirs (6) is related to differences in whole-blood RNA extraction methods due to a presumed difference in pH of the phenol-guanidinium. The Trizol U.S. patent documents state that the pH of Trizol is 4.0, the exact pH used in their method