We read with interest the article by Sebastiani and colleagues regarding the use of SAFE biopsy in patients with chronic hepatitis C infection.1 Despite being a large and well-conducted study, there remain problems with the algorithm that may limit its use. The requirements of a screening test are that it must be accurate, cost-effective, and easy to use. Furthermore, the purpose and use of such a test must be well defined. The purpose of the algorithm proposed is first to identify patients with significant fibrosis (i.e., METAVIR ≥ F2) and second to determine the probability that a patient has cirrhosis, and is therefore in need of long-term outpatient follow-up for the complications of cirrhosis. In the United Kingdom, the stage of liver fibrosis has not been a prerequisite for chronic hepatitis C treatment since 2006, obviating the need for liver biopsy for this indication.2 Patients only need to be positive for hepatitis C virus RNA to be a candidate for treatment. It might be argued that patients with genotype 1, who have a less favorable response to current treatment, may choose to defer treatment if lower stage disease (METAVIR F0-1) can be demonstrated, the intention being to postpone treatment until more efficacious and less toxic treatments become available. This is something we encounter weekly in our clinics. Based on this we would argue that any treatment algorithm should include genotype and patient age, because this has a significant impact on the decision to initiate therapy.
In effect, SAFE biopsy consists of two screening tests prior to biopsy. This increases the complexity of how to interpret the results. In addition, patients and clinicians like to have a test result that gives an estimate of fibrosis severity at the time they are seen and that is easy to calculate. This is the advantage of techniques such as transient elastography3 and other noninvasive tests such as aspartate aminotransferase-to-platelet ratio index (APRI)4 and Kings score.5 In order for this to be achieved, blood tests would be required prior to assessment. Finally, the percentage of patients with cirrhosis in this cohort of patients was low. Any noninvasive test compared to liver biopsy should consider the frequency of particular stages of fibrosis to reduce error using the method described by Poynard and colleagues.6 This low level of cirrhosis may account for why the area under the receiver operating characteristic curves were lower for APRI than in other published series.4, 5, 7