The article by Carrion et al.1 recently published in Liver Transplantation is the first study that has evaluated transient elastography (Fibroscan) in relation to hepatic venous pressure gradient (HVPG) measurements, and liver histology in patients with recurrent hepatitis C (HCV) infection after liver transplantation. There were 124 patients who underwent 169 liver biopsies (LB) between July 2004 and October 2005. Transient elastography was found to correlate well with HVPG (Pearson coefficient, 0.84; P < 0.001), whereas a cutoff value of ≥8.74 kPa had excellent predictive value for portal hypertension (i.e., HVPG ≥6 mm Hg). In addition, a cutoff value of ≥8.5 kPa had excellent performance in determining significant fibrosis (stage F2-F4).
However, the assessment of transient elastography, and indeed HVPG, depends on the quality of the LB specimens, which is the gold standard for both of these measurements. This is where we think that the study has some potential important limitations. Although the needle size was stated (14 gauge for percutaneous LB [PLB] and 16 gauge for transjugular LB [TJLB]), there were no details given for the sample's length, number of portal tracts obtained, or fragmentation. The authors do state that “optimal” biopsy samples were obtained (without defining what they mean), but although 39% of liver biopsies were PLB and 61% were TJLB. In a previous article,2 the same authors evaluated patients transplanted for HCV cirrhosis between October 2000 and May 2003. Carrion et al. thus either report the same patients, or, if further liver transplant recipients were assessed, the 2 cohorts share most patients in common. The average length of TJLB was only 10 mm and 15 mm for PLB, with an average of 7 and 9 (summation of complete and partial) portal tracts, respectively.2 Blasco et al.2 could not correlate HVPG with the TJLB because of the inferior quality of the LB, but these represent 61% of the biopsies in the current study.1 We submit that if they found their TJLB suboptimal to correlate histology with HVPG in the previous study,2 then they are suboptimal in comparison with HVPG in this study1 and with transient elastography.
This is not to say that TJLB cannot provide sufficiently good, even optimal, LB specimens. We have formally evaluated the quality of a standardized TJLB by using a 19-gauge Trucut needle and 3 passes3 and found that the quality was similar to PLB, with a medium length of 22 mm and complete portal tracts of 8 and partial portal tracts of 5 (13 total). However, our systematic review has suggested that most published series of PLB and TJLB are suboptimal.4 Specifically, a single-pass TJLB is insufficient; at least 3 passes are needed.4 There may also be differences between TJLB performed with Menghini needles (used by Blasco et al.2 and, we assume, in any added patients that may have undergone biopsy in the study by Carrion et al.1) compared with Trucut needles.
Given that current standards suggest optimal grading and staging in chronic HCV infection should be performed with LB samples 20-25 mm in length and/or containing ≥11 complete portal tracts,5 then we believe transient elastography should be assessed against biopsy procedures that nearly reach or fulfill these standards. Clearly, neither HVPG nor transient elastography had an adequate histological reference standard in either study.1, 2
Moreover, given that transient elastography cannot be a surrogate for other histological features such as inflammation and steatosis, which are reported to have good predictive value for the progression of fibrosis,6, 7 LB will still be needed to assess recurrent HCV disease after liver transplantation. Because LB will need to be performed, and because measurement of HVPG has independent prognostic value,2 the transjugular route offers the most efficient route to perform both biopsy and HVPG measurement at the same time, and it allows multiple passes to enable better, often optimal liver samples to be obtained.3 This will provide the basis for an accurate assessment of transient elastography or of other noninvasive markers of fibrosis and of HVPG itself.