We are delighted that Lackner et al. provided external validation for our two models for predicting cirrhosis among patients with hepatitis C. Platelet count carried the most weight in both of these models; therefore, it is not surprising that Lackner et al. found that accuracy of platelet count alone approximated the more complex models. Our experience with both treatment-naïve patients and nonresponders with advanced liver disease indicate that platelet count alone was not as accurate as APRI or model 3, because some patients who had histological cirrhosis had normal platelet count, whereas other patients with low platelet count did not have cirrhosis on biopsy. We agree that an ideal model should perform equally well in predicting and excluding cirrhosis, but the latter itself has clinical relevance as indicated by Lackner et al.
Anna S.F. Lok*, * University of Michigan, Ann Arbor, MI.