Liver biopsy is essential in the follow-up of HCV-infected liver transplant recipients. The aim of this study was to prospectively compare percutaneous (PLB) versus transjugular liver biopsy (TLB) in the assessment of liver damage. We also explored the diagnostic value of hepatic venous pressure gradient (HVPG) to identify patients at risk of severe HCV disease recurrence after liver transplantation (LT). One hundred sixteen paired PLB and TLB (with HVPG measurement) were performed 3 or 12 months after LT in 80 patients. Concordance for necroinflammation and fibrosis was fair or good, particularly 1 year after LT (kappa ≥ 0.6). At this point, a significant positive association was seen between the median HVPG and the fibrosis stage (2.5 mm Hg for F0; 5 mm Hg for F1, 6 mm Hg for F2, and 11.5 mm Hg for F3; Kruscal-Wallis < 0.001). Despite this strong association, portal hypertension (HVPG ≥ 6 mm Hg) was detected in 1 (5%) of 22, 4 (16%) of 25, and 6 (60%) of 10 patients with fibrosis stages 0, 1, and 2, respectively. After a median follow-up of 38 months, clinical decompensation occurred in 15 (19%) of 80 patients. Although the presence of significant fibrosis (F2-F3) 1 year after transplantation was good to predict clinical decompensation (AUC: 0.80), an HVPG of 6 mm Hg or greater was extremely accurate at identifying patients at risk of disease progression (AUC: 0.96). In conclusion, HVPG determination is a valuable tool for follow-up in patients with HCV recurrence after LT. (HEPATOLOGY 2006;43:492–499.)
Chronic hepatitis C virus (HCV) infection leading to liver cirrhosis and hepatocellular carcinoma is the main indication for liver transplantation (LT) in Western countries and Japan.1 Regretfully, recurrence of HCV infection is universal after LT,2 and disease progression is significantly faster in immunosuppressed than in immunocompetent individuals. In liver transplant recipients, chronic HCV infection may lead to cirrhosis in as much as 30% of individuals only 5 years after LT.3–5 Once liver cirrhosis is established, the cumulative probability of developing clinical decompensation is close to 50% 1 year after diagnosis and, more importantly, survival after decompensation is extremely short.6 As a result of this accelerated course of HCV infection, long-term graft and patient survival are significantly reduced in patients undergoing LT for HCV-related cirrhosis compared with other groups.7 Frequent liver biopsies are essential to monitor HCV-induced liver damage and are part of the routine follow-up of HCV-infected liver transplant recipients. Early histological damage after transplantation correlates with long-term outcome; in fact, the presence of significant liver fibrosis in 1-year liver biopsies identifies patients at high risk of graft loss.3, 8 In addition, assessment of liver damage is relevant to adopt therapeutic decisions, particularly because of the low efficacy and high incidence of adverse events of current antiviral therapy in this group of patients.3, 9, 10
Sampling variability is a limitation of liver biopsy for the assessment of fibrosis. Needle biopsy specimens represent an extremely small part of the liver and for this reason sample errors are frequent.11, 12 More importantly, underestimation of liver damage might become a relevant issue in individuals with rapid disease progression, such as HCV-infected liver transplant recipients.
Usually, liver biopsy is performed percutaneously, guided by ultrasonography (US). The latter procedure allows the evaluation of vessel permeability and the presence of abdominal collections during the first weeks after transplantation. Transjugular liver biopsy (TLB) is indicated in individuals with coagulation disorders (or ascites) in whom the percutaneous approach is contraindicated. Although tissue samples obtained by this approach are usually smaller than those obtained by the percutanous route, catheterization of the hepatic veins allows the measurement of hepatic venous pressure gradient (HVPG).13 HVPG accurately reflects portal pressure in HCV-related cirrhosis14 and if adequately measured has a very low variability.13 In the absence of significant liver disease, HVPG does not exceed 5 mm Hg.13 HVPG reflects the interaction between hepatic vascular resistance and blood flow and, as such, is thought to closely indicate disease severity. Actually, it has been shown to correlate with survival, decompensation, and development of collaterals.15, 16 These considerations suggest that HVPG could be used as a quantitative marker of disease progression in patients with chronic hepatitis C.17 This would be particularly relevant in liver transplant recipients with HCV infection recurrence.
The aim of this study was to prospectively compare percutaneous liver biopsy (PLB) versus TLB to assess liver damage in HCV-infected liver transplant recipients, as well as to explore the diagnostic value of HVPG to identify patients at risk of severe HCV disease recurrence.
HCV, hepatitis C virus; LT, liver transplantation; US, ultrasonography; TLB, transjugular liver biopsy; HVPG, hepatic venous pressure gradient; PLB, percutaneous liver biopsy; ROC, receiver operator characteristics; AUC, area under the curve.
Patients and Methods
HCV-infected patients who underwent primary LT for end-stage cirrhosis or hepatocellular carcinoma between October 2000 and May 2003 were included in the study. In our Institution, PLB are performed on a protocol basis 3, 12, 24, 36, and 60 months after LT in HCV-infected patients. For this study, a TLB with HVPG measurement was performed 24 hours after the protocol PLB. Exclusion criteria were (1) combined kidney and liver transplantation; (2) co-infection with the hepatitis B virus; (3) survival shorter than 3 months after transplantation; (4) rejection or suspicion of rejection; (5) biliary tract complications; (6) clinical decompensation (ascites, hepatic encephalopathy); (7) antiviral therapy at time of biopsy; (8) retransplantation. After a detailed discussion of the protocol, including the potential risks of a TLB, patients who gave their written informed consent were included in the study. The protocol had been previously approved by the Investigation and Ethics Committee of the Hospital Clinic of Barcelona.
During hospital admission, patients were managed according to previously published protocols.4 In brief, induction immunosuppression was cyclosporine A or tacrolimus and prednisone. Mycophenolate mofetil was added in patients who required cyclosporine or tacrolimus dose reduction or discontinuation. Immunosuppression therapy was recorded throughout the study. Acute rejection episodes were documented by liver histology and treated with steroid boluses if moderate or severe. After discharge, patients were visited at the outpatient clinic, monthly for the first 3 months with complete record of clinical and analytical variables (including viral load), and every 2 months thereafter. Patients were followed until June 2005 or until death, or retransplantation. Follow-up liver biopsies and liver hemodynamic studies, analytical data, endoscopic and imaging studies, as well as the occurrence of clinical decompensation, were recorded prospectively during follow-up.
PLB were performed under local anesthesia and US guidance with a Tru-Cut 14G needle by expert radiologists. Thirty minutes before the procedure and after an overnight fast, patients received 5 mg diazepam and 5 mg atropin.
Twenty-four hours after the PLB and after an overnight fast, a TLB with HVPG measurement was performed at the Hepatic Hemodynamics Laboratory. Under local anesthesia, a venous introducer was placed in the right internal jugular vein by the Seldinger technique. Under fluoroscopy, a 7F balloon-tipped catheter (Boston Scientific Medi-Tech, Natick, MA) was guided into the main right or middle hepatic vein for measurements of wedged and free hepatic venous pressures. Adequacy of occlusion was checked by injection of a small amount of radiologic contrast medium. The portal pressure gradient was measured as the HVPG, the difference between wedged and free hepatic venous pressures, as previously described.18 All measurements were performed in triplicate, and permanent tracings were obtained on a multichannel recorder (Marquette Electronics, Milwaukee, WI) and read by an experienced investigator unaware of the clinical conditions of the patient. Later the catheter was exchanged through a guide for a 10F catheter. To obtain a liver biopsy, a 16G needle (CooK, Bjaverskov, Denmark) connected to a 20-mL syringe was placed until the distal part of the catheter. Thereafter the needle was introduced 2 to 4 cm within the liver tissue while aspirating. This procedure was repeated (median, 2 times) until an optimal tissue sample was obtained. Patients were sedated with midazolam during the entire procedure, using a dose (0.02 mg/kg intravenously) that has been shown not to influence hepatic hemodynamics.19
Processing of Liver Biopsy Specimens.
Samples were processed at the Pathology Department and stained with hematoxilin-eosin and Masson's trichromic. All histological samples were scored independently by two pathologists (M.B., R.M.) who knew neither the clinical status of the patient nor whether biopsy samples belonged to the same patient. Special emphasis was taken to exclude other causes that might explain portal hypertension, including diseases affecting the centrilobular vein. Necroinflammatory activity and fibrosis stage were scored using the Scheuer classification.20, 21 In addition, the length, width and the number of portal tracts of each specimen were recorded.
Quantitative variables are expressed as median (range). Differences between qualitative variables were assessed by the Fisher exact test. Differences between quantitative variables were analyzed by a non-parametric test (Mann-Whitney or Kruskal-Wallis for unpaired samples, Wilcoxon's for paired samples). Concordance was assessed by kappa statistics (kappa ≤ 0.4 low, kappa 0.41-0.6 fair; kappa > 0.6 good). Cumulative probability curves of clinical decompensation according to Kaplan-Meier method were compared by the Mantel-Cox test. The diagnostic value of liver fibrosis and HVPG 1 year after LT to predict clinical decompensation was assessed by calculating the areas under the receiver operator characteristic (ROC) curves. An area under the curve (AUC) of 1.0 is characteristic of an ideal test, whereas 0.5 indicates a test of no diagnostic value. The software used for statistical analysis was SPSS 11.0 (SPSS Inc., Chicago, IL).
Between October 2000 and May 2003, 135 liver transplantations were performed in 125 patients with HCV-related cirrhosis. From this cohort, the final number of individuals included in the study was 80 (Fig. 1). The baseline characteristics of these 80 patients are shown in Table 1.
Table 1. Baseline Characteristics of the 80 Patients Included in the Study
The total number of paired liver specimens was 116 (51 at month 3, and 65 at month 12 after LT) (Fig. 1). As expected, specimens obtained by TLB were significantly smaller and had a lower number of portal tracts than samples obtained by the percutaneous route (10 mm vs. 15 mm of length, P < .01; 7 vs. 9, P < .01 ; respectively).
Evaluation of Necroinflammatory Activity and Fibrosis Stage.
Mild acute hepatitis was the most common finding in biopsies performed 3 months after LT, whereas chronic hepatitis was almost universal 12 months after LT. Histological findings were quite similar between the 51 paired transjugular and percutaneous specimens obtained 3 months after transplantation, and the concordance was fair both for lobular activity (kappa = 0.54) and portal-periportal activity (kappa = 0.5 ). Regarding specimens obtained 12 months after LT, concordance was good for lobular activity (kappa = 0.68) and fibrosis stage (kappa = 0.6) and fair for portal-periportal activity (kappa = 0.5). Results for all specimens are shown in Table 2.
Table 2. Evaluation of Necroinflammatory Activity and Fibrosis in 116 Paired Specimens: Comparison Between Percutaneous and Transjugular Liver Biopsy
Transjugular Liver Biopsy
Percutaneous liver biopsy
Association Between Liver Fibrosis and HVPG.
HVPG and liver biopsy specimens were available in 51 patients 3 months after LT and in 65 patients 1 year after LT. Because of the larger size of biopsies obtained by the percutaneous procedure, fibrosis stage scored in PLB was used for statistical analyses. Because the small number of patients with advanced fibrosis or abnormal HVPG 3 months after transplantation, the statistical analysis was not performed at this time point.
One year after liver transplantation, 18 patients had significant liver fibrosis (F ≥ 2) and 18 had portal hypertension (HVPG ≥ 6) (Fig. 2). There was a significant positive relationship between the fibrosis stage and HVPG (Kruscal-Wallis < 0.001) (Fig. 2). Despite this strong association, portal hypertension (HVPG ≥ 6 mm Hg) was detected in one (5%) of 22, 4 (16%) of 25, and 6 (60%) of 10 individuals with fibrosis stages 0, 1, and 2, respectively. Importantly, HVPG was 12 mmHg of greater in five of these individuals (Fig. 2). Because sinusoidal portal hypertension is not supposed to develop in the absence of significant liver fibrosis, we recorded all relevant clinical, laboratory, histological, and endoscopic follow-up data of patients with HVPG 6 mm Hg or greater (Table 3). In all patients, causes of portal hypertension different from HCV infection recurrence were reasonably excluded by ultrasound examination, hepatic vein catheterization, and liver biopsy. The rate of previous rejection episodes was not different between patients with portal hypertension (3 of 18, 17%) and those without portal hypertension (10 of 47, 21%) (P = .68). Importantly, histological evidences of rejection were carefully investigated and not detected in any of the liver biopsies of patients included in the study.
Table 3. Clinical, Analytical, and Histological Follow-up of Patients With Elevated HVPG After One Year Following LT
HVPG (w, f) (12 m)
Fibrosis (12 m)
HVPG (w, f) (24 m)
Fibrosis (24 m)
Proteins Ascitic Fluid (g/L)
Bilirubin, Albumin, Prothrombin (at Time of Decompensation)
Because most patients had follow-up liver biopsies (57 of 65), we analyzed whether progression of fibrosis (increase in at least one fibrosis stage) was different depending on the HVPG value 1 year after transplantation. Among 18 patients with portal hypertension 1 year after LT, the follow-up liver biopsy (n = 17) demonstrated fibrosis stage F2 in 4 (23%), F3 in 3 (18%), and F4 in 10 (59%). Among 47 patients with normal portal pressure 1 year after LT, the follow-up liver biopsy (n = 40) demonstrated fibrosis stage F0 in 9 (22%), F1 in 20 (50%), F2 in 8 (20%), and F3 in only 3 (8%). Therefore, follow-up liver biopsy demonstrated progression of fibrosis in 13 (76%) of 17 individuals who had portal hypertension 1 year after LT compared with 19 (47%) of 40 with normal HVPG (P = .044).
Liver Fibrosis and HVPG as Predictors of Clinical Decompensation.
After a median of 38 months after transplantation (range, 20–61), clinical decompensation caused by progression of HCV disease recurrence occurred in 15 (19%) of 80 patients. Administration of antiviral therapy was significantly more frequent in patients who presented clinical decompensation (9 of 15, 60%) compared with patients who did not present decompensation (13 of 65, 20%) (P = .004). The figure was similar for the 65 patients with paired 1-year liver biopsies (7 of 13, 54% vs. 11 of 52, 21%) (P = .018) This can be explained by the fact that antiviral therapy is indicated preferentially in individuals with severe HCV infection recurrence.
ROC curve analysis showed that fibrosis stage (scored in percutaneous liver biopsies) 3 months after LT was not accurate at identifying patients at risk of clinical decompensation (AUC: 0.67). This is probably explained by the mild histological findings at this time point. Clinical decompensation occurred in 2 of 4 patients with liver fibrosis stage 2 or greater versus 8 of 47 with fibrosis stage F0 to F1 (log-rank = 0.053). HVPG measured 3 months after LT was slightly better than the fibrosis stage at identifying patients at risk of clinical decompensation (AUC: 0.76). Among eight patients with HVPG of 6 mm Hg or greater, clinical decompensation developed in four, whereas decompensation occurred only in 6 of 43 patients with normal HVPG at this time point (log-rank = 0.025).
Clinical decompensation occurred in 13 (20%) of 65 patients with paired 1-year liver biopsies (median follow-up after the procedure, 28 months). Decompensation was documented in 4 of 47 individuals with fibrosis stages 0 to 1 versus 9 of 18 patients with fibrosis stages 2 to 3 (log-rank < 0.01) (Fig. 3). The value of significant liver fibrosis to predict clinical decompensation improved considerably 1 year after transplantation (AUC: 0.80) (Fig. 4). Regarding HVPG measurements, clinical decompensation occurred in only 1 of 47 patients with HVPG less than 6 mm Hg versus 12 of 18 patients with HVPG of 6 mm Hg or higher (log-rank < 0.01) (Fig. 3). Importantly, HVPG was extremely accurate at identifying patients at risk of disease progression (AUC: 0.96) (Fig. 4). The best HVPG cut-off value was 6 mm Hg or greater, which is used to define portal hypertension. As shown in Table 4, the accuracy of HVPG was better than liver fibrosis to predict clinical decompensation.
Table 4. Diagnostic Accuracy of Liver Fibrosis and HVPG 1 Year After Liver Transplantation to Predict Clinical Decompensation
NOTE. Liver fibrosis was scored in specimens obtained by percutaneous liver biopsy.
< 6 mmHg
≥ 6 mm Hg
Fibrosis ≥ 2
HVPG ≥ 6 mm Hg
Complications Related to the Procedure.
Complications associated with percutaneous or transjugular liver biopsies were recorded prospectively. Hemoperitoneum related to PLB occurred in 2 (2%) of the 116 procedures; in both cases diagnosis was made by US after clinical suspicion and resolved spontaneously after bed rest. Fever after TLB was documented in 2 (2%) of 116 procedures; in both cases it was transient, and blood cultures were negative.
Costs of the Procedures.
The cost of a liver hemodynamic study relative to the cost of a percutaneous liver biopsy is practically identical (1.05). If a transjugular liver biopsy is performed, the procedure is more expensive (1.45) relative to a PLB.
HCV disease recurrence is the main problem of liver transplant programs. Several studies have demonstrated that progression of fibrosis is significantly faster in immunosuppressed than immunocompetent individuals.4–6, 22–24 For this reason, frequent liver biopsies are part of the routine follow-up of HCV-infected patients after LT.3, 4, 8, 25, 26 Most centers perform liver biopsy by the percutaneous route under ultrasonographic control. Besides the advantages of US (particularly during the first weeks after LT), liver biopsy allows a semiquantitative evaluation of necroinflammatory changes and fibrosis, as well as identification of additional pathological conditions (such as rejection). The presence of significant fibrosis 1 year after liver transplantation appears to identify a significant proportion of patients at high risk of graft loss in whom antiviral therapy is clearly indicated.8 Sampling variability, however, is a significant limitation in the assessment of liver fibrosis with liver biopsy. Underestimation of disease severity might have serious consequences in patients with rapid disease progression.
We found a significant positive association between the fibrosis stage scored in liver biopsies performed 1 year after transplantation and the HVPG measurement at the same time. Despite this strong association, a significant proportion of patients without a histological diagnosis of cirrhosis in the liver biopsy had portal hypertension. Although the latter finding was particularly frequent for individuals with fibrosis stages 2 and 3, some patients without fibrosis or mild fibrosis (without architectural distortion) had an HVPG value of 6 mm Hg or greater. This is well exemplified by three patients with fibrosis stage 0 and 1 with an HVPG of 12 mm Hg or greater, which is the threshold for clinical complications. Liver hemodynamics and follow-up data (including repeated liver biopsies and HVPG measurements, laboratory, and clinical data) reasonably excluded other causes of portal hypertension and reinforced the hypothesis that HCV recurrence was the cause of increased HVPG. However, the presence of portal hypertension in the absence of significant liver fibrosis is difficult to explain. Because follow-up liver biopsies indicated cirrhosis or bridging fibrosis in patients with discrepant results, sampling errors may have occurred in some of these cases. This is further supported by our finding that individuals with elevated HVPG (≥6 mm Hg) 1 year after transplantation are at high risk of clinical decompensation (see below). Apart from sampling errors, another possible explanation is the different pattern of fibrosis deposition in liver transplant recipients compared with immunocompetent patients. We have observed that in liver transplant recipients with HCV recurrence the presence of fibrosis within sinusoids is relatively frequent in the absence of alcohol intake (M. Bruguera, unpublished observation). In fact, liver biopsy specimens of one individual with mild fibrosis and portal hypertension revealed the presence of significant pericellular fibrosis, expanding along the sinusoid. Preferred deposition of collagen within liver sinusoids (with fibrosing cholestatic hepatitis being its maximum expression) may cause portal hypertension in the absence of fibrous septa. Current liver fibrosis staging systems do not take into consideration this possibility and may be not entirely adequate to score fibrosis in transplant recipients. Our results strengthen the value of HVPG measurement as a dynamic test to assess the progression of liver diseases in the pre-cirrhotic stage and are especially relevant from a clinical point of view, because in patients with rapid disease progression delaying antiviral therapy might have serious consequences.
In fact, the most relevant finding of our study was the excellent predictive value of HVPG at identifying patients at risk of clinical decompensation. Clinical decompensation occurred in only 1 (2%) of 47 patients with normal HVPG versus 12 (67%) of 18 patients with abnormal HVPG 1 year after LT. ROC analysis demonstrated that HVPG was more accurate than fibrosis staging at identifying patients with risk of severe disease recurrence, which has been clearly associated with graft loss and death.6
HVPG measurement has several advantages compared with liver biopsy. The most important is that balloon-catheter measurements of HVPG are not subject to sampling errors.13 Moreover, HVPG measurements are very reproducible in experienced hands.13, 14, 17 The risks of the procedure are very low, particularly if no liver biopsy is obtained. In case of liver biopsy, the risks are not higher by the transjugular than the by the percutaneous route.27, 28 Technical success in experienced centers is higher than 95% and, as shown in this and other studies, specimens are adequate for histology in more than 90% of cases.27, 28 Despite the smaller size of specimens obtained by the transjugular route, the diagnostic agreement between percutaneous and transjugular liver biopsy was good, particularly in samples obtained 1 year after LT. Although underestimation of liver damage occurs in small specimens,11 the level of experience of the pathologist is also relevant for diagnostic accuracy.12
HVPG measurement (and TLB) has some limitations. First, it is an invasive and expensive procedure, which is of concern, particularly if it needs to be repeated during follow-up. Although liver biopsy remains as the gold standard for diagnosis of liver fibrosis, several scores based on easily available biochemical or hematological parameters have shown a significant value in assessing the fibrosis stage in patients with chronic hepatitis C.29–32 Although virtually no reports exist that analyze the value of laboratory data to assess liver fibrosis in transplant recipients,33 in the next few years serum fibrosis markers or other non invasive tests (such as transient elastography34) will likely prove useful to assess HCV disease severity in transplant recipients. Another limitation of HVPG measurement is that it is not performed in all medical centers. However, liver transplant programs are settled in referral centers, where facilities and staff experienced in liver hemodynamics are usually present.
In summary, our data suggest HVPG determination is an extremely valuable tool to follow-up patients with HCV recurrence after liver transplantation. The finding of an increased HVPG 1 year after LT identifies with high accuracy patients with severe HCV recurrence or at high risk of disease progression. Our findings further suggest that HVPG would be very useful to monitor the response of these patients to antiviral therapy.