We thank Dai et al. for their comments regarding our study. May we first remind them that this study mainly showed a strong relationship between early ribavirin exposure measured at Day 0 (D0) and sustained virological response (SVR). Abbreviated area under the concentration curve (AUC0-4h) at this date was as strongly linked to SVR as rapid virological response (RVR) (odds ratio 7.0 and 6.7, respectively). Studying the relationship between D0 AUC and RVR or early virological response (EVR) was only a secondary objective. We agree that our definition of RVR (> 2 log drop or undetectable viral load at Week 4 [W4]) is wider than that of the Asian consensus (negative polymerase chain reaction at W4).1 However, other authors such as Ferenci et al. used the same definition as ours in their interesting retrospective analysis of 465 patients treated with peginterferon alfa-2a and ribavirin, and observed a positive predictive value (PPV) of 75% and a negative predictive value (NPV) of 74%.2 Moreover, when our study was performed, there was no published definition of RVR using the highly sensitive quantitative real-time polymerase chain reaction (PCR) assay HCV Ampliprep TaqMan (cut-off 15 UI/mL). It is noticeable that, using this technique, viral load remained detectable at W4 for most of our study patients. Consequently, the definition of RVR based on undetectable viral RNA at W4 did not seem appropriate in this case. We therefore chose an RVR definition that remained reliable whatever the sensitivity of the assay (that is, > 2 log drop or undetectable viral load).
Negative PCR at W12, known in the literature as “complete early virological” response is now a useful tool for tailoring treatment duration in patients infected with genotype 1.3 Its PPV and NPV were 68% and 83%, respectively, in a retrospective analysis from two large multinational phase III studies of 569 patients infected with genotype 1.4
The definitions we used for RVR and EVR in our article seemed to be good compromises between NPV and PPV and provided us with a more precise global predictor of our final endpoint (SVR) than the classical ones, which was important for the subsequent multivariate statistical analysis.
We agree with Dai et al. that the percentage of RVR and EVR is highly dependent on the definition used, the sensitivity of the technique employed (Real time PCR versus Cobas Amplicor PCR assay [cut-off 50 UI/mL]) and the populations studied. However, it had no impact on the high predictive value of very early ribavirin exposure in term of SVR. A randomized trial is of course necessary to validate the clinical impact of ribavirin individual dose adjustment based on D0 abbreviated AUC, using the target value we proposed, first in a population comparable to the one studied in the exploratory trial and then of course in others.