Hepatocellular Carcinoma Screening in Patients Treated for Hepatitis C

The American Journal of Medicine
February 2018 Volume 131, Issue 2, Page e79

Letter to the editor
Hepatocellular Carcinoma Screening in Patients Treated for Hepatitis C
Harsh N. Patel, MD, Jason R. Stibbe, MD, Meghana Vellanki, MD, Harold Paul, MD

Article reference -  Hepatocellular Carcinoma Screening Associated with Early Tumor Detection and Improved Survival Among Patients with Cirrhosis in the US

http://www.amjmed.com/article/S0002-9343(17)30128-6/fulltext

Fla DOI: http://dx.doi.org/10.1016/j.amjmed.2017.09.036

To the Editor:
We have read your article by Singal et al1 with great interest, which discussed hepatocellular carcinoma screening being associated with early tumor detection and improved survival among patients with cirrhosis in the United States. The study clearly highlighted the importance of performing hepatocellular carcinoma screening in patients with cirrhosis; however, what is of concern is that more than half of the cases of hepatocellular carcinoma were found secondary to symptoms or incidentally, rather than through the hepatocellular carcinoma screening algorithm. Although adherence to screening was improved with subspecialty referral, patients are largely cared for by primary care physicians—highlighting a critical area in need of intervention.

In this study, the most common cause for cirrhosis in the population studied was hepatitis C virus at 57.2%; since 2013 there has been a shift in the treatment of hepatitis C viral therapy with direct-acting antivirals. Historically, interferon-based regimens have been the mainstay of anti–hepatitis C virus therapy, yielding hepatitis C virus cure or sustained virologic response in approximately 50% of patients.2 Recently developed direct-acting antivirals, which directly target the viral protease, polymerase, or nonstructural proteins, have revolutionized interferon-free regimens leading to an improved rate of achieving sustained virologic response, approaching or surpassing 90%.3

Although rates of hepatitis C virus–associated hepatocellular carcinoma will decrease significantly after the widespread adoption of direct-acting antivirals, there remains a persistent risk for hepatocellular carcinoma among patients with advanced fibrosis who have achieved sustained virologic response.4 The most well-established risk factor for developing hepatocellular carcinoma after a sustained virologic response in patients with hepatitis C virus is advanced fibrosis or cirrhosis.5 In a retrospective cohort study of 562 patients with sustained virologic response, the 5-year survival rate for those who received regular hepatocellular carcinoma screening (ultrasonography at least every 6 months) was 93%, whereas it was 60% for those who did not.4 As such, suggestion would be that individuals with hepatitis C virus with advanced fibrosis treated with anti–hepatitis C virus therapy and who have a sustained virologic response should continue to be screened regularly for hepatocellular carcinoma because the risk of disease progression, including hepatocellular carcinoma development, is not completely eliminated after sustained virologic response. Guidance should be provided to primary care providers to ensure the increasing population of patients with sustained virologic response after treatment for hepatitis C virus are maintained in the patient population needing screening.
http://www.amjmed.com/article/S0002-9343(17)31022-7/fulltext?rss=yes

The Reply

Gregory Seymann, MD, SFHM Correspondence information about the author MD, SFHM Gregory Seymann Email the author MD, SFHM Gregory Seymann , Robert El-Kareh, MD, MS, MPH, Gabrielle Schaefer, MD, Jennifer Quartarolo, MD

DOI: http://dx.doi.org/10.1016/j.amjmed.2017.09.037

The authors have reviewed the thoughtful commentary by Ayubi and Safiri on our article and appreciate the interest in our work.1

Ayubi and Safiri contend that our attempt to validate the Yale New Haven Readmission Risk Score in a different population of patients with pneumonia is limited by a lack of methodological transportability, given we incorporated new variables of functional status and social support that were not tested in the original study. We refer the reader to Table 2, in which we report modeling of the variables in the original Yale New Haven Readmission Risk Score, as well as modeling with the new variables. They also raise the concern for need of cross-validation in our methodology. Had we found significant contribution by our new predictors, then we would have pursued cross-validation to ensure that those results were not due to overfitting our sample. However, because our overall results were consistent with the larger study that developed the Yale New Haven Readmission Risk Score and our new predictors were not found to have a significant contribution, we did not pursue cross-validation.

They note concerns about selection bias and characterize our sample as a convenience sample. As is detailed in the “Methods” section, we included all patients hospitalized with pneumonia at our institution over a 1-year period. We respectfully disagree that a selection bias would be introduced by using a complete sample, although the limitations of a single-center sample, addressed in our discussion, still apply.

We agree with the authors that the addition of functional status and social support to the model did not produce clinically significant improvement in the predictive power of the model. This is pointed out in the “Discussion” and “Conclusions,” although we acknowledged that more robust markers of these variables may yield a more meaningful comparison.

Regarding the authors' question about our model-building methodology, we chose the predictors of the model a priori and developed our logistic regression model on the basis of those selections. We did not use other techniques to add or remove any predictors.
http://www.amjmed.com/article/S0002-9343(17)31023-9/fulltext