点击显示 收起
Ordway Research Institute, Albany, New York
I would like to thank Shenoy et al. [1] for their commentary on our recent article [2]. I agree that baseline viral load and early response are associated with the probability of a sustained viral response (SVR). The earlier work mentioned by Shenoy et al. [3] was cited in our article.
What is important is that hepatitis C virus (HCV) genotype 1 infections have a much lower response rate than HCV nongenotype 1 infections, even when the variables pointed out by Shenoy et al. are accounted for. The question raised is whether the duration of therapy would affect the SVR; given the seminal study of Poynard et al. [4], it is our contention this is a reasonable hypothesis to address. Examination of a large database, with model building in 1 group (771 patients) and model validation in another group, which was not used for model building (298 patients), demonstrated positive and negative predictive values, as well as a sensitivity and specificity, all of which exceeded 91% in the validation step. This clearly does not prove the point and, as was stated in our article, merely points out the need for further investigation in this area. A prospective, randomized trial would be required to prove the hypothesis, as we stated.
In the published model, once the duration of negativity was entered into the model, the backward elimination procedure took one of the factors cited by Shenoy et al. (baseline viral load) out of the model. The database did not allow us to calculate the early viral response (EVR). The deletion of baseline viral load is understandable, because this factor drives an earlier time to negativity and, with a fixed duration of therapy (48 weeks in the trial), this leads to a longer duration of negativity. Certainly, starting off with a lower baseline viral load (or having a good early decline, but we could not examine this) would lead to an earlier time to negativity and longer durations of negativity.
Furthermore, I believe that Shenoy et al. have misinterpreted the entry in table 2 of our article [2]. If one looks at the second column of table 2, under "Virologic" variables, all of the probabilities are approximately equal (all <9 × 10-7). The third column is the McFadden's 2, which is considerably larger for clearance duration. I believe that this is the source of the confusion. This number is the logistic regression analogue of the r2 value of linear regressionthe amount of variance explained by the variable. In this case, higher is better for McFadden's 2. Clearance duration was most explanatory of the variance in attaining the defined outcome, SVR. It is then not surprising that baseline viral load was removed from the final model in the backward elimination procedure.
The point of our article was not to denigrate the effect of baseline viral load or to ignore it (or EVR, had we been able to examine it). Rather, it was a hypothesis-generating article indicating that an exploration of the duration of treatment for patients infected with HCV genotype 1 might be fruitful in raising the probability of attaining a SVR. Clearly, such a strategy would need to balance the possible benefits with the risks of extra toxicity and the possibility of poor adherence to the regimen after a certain point. Extending the duration of treatment from 24 to 48 weeks [4] also made it necessary to explicitly balance the risk versus benefit, but it was clearly shown to be beneficial, on balance. Only when a randomized trial is performed to examine the hypothesis generated by our analysis can we explicitly make a risk-benefit judgment for patients infected with HCV genotype 1. EVR and baseline viral load may play a role in that judgment.
References
1. Shenoy M, Levin J, Said A, Striker R. Clearance duration as a predictor of sustained viral response in patients with chronic hepatitis C virus genotype 1 . J Infect Dis 2005; 191:19935 (in this issue). First citation in article
2. Drusano GL, Preston SL. A 48-week duration of therapy with pegylated interferon 2b plus ribavirin may be too short to maximize long-term response for patients infected with genotype-1 hepatitis C virus. J Infect Dis 2004; 189:96470. First citation in article
3. Manns MP, McHutchinson JG, Gordon SC, et al. Peginterferon alfa-2b plus ribavirin for initial treatment of chronic hepatitis C: a randomised trial. Lancet 2001; 358:95865. First citation in article
4. Poynard T, Marcellin P, Lee SS, et al. Randomized trial of interferon 2b plus ribavirin for 48 weeks or 24 weeks versus interferon 2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. International Hepatitis Interventional Therapy Group. Lancet 1998; 352:142632. First citation in article