If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Hospital Clínic, IDIBAPS, Universitat de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Hospital Clínic, IDIBAPS, Universitat de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Hospital Clínic, IDIBAPS, Universitat de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, SpainCentro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, Madrid, Spain
Liver stiffness improves after antiviral therapy in most patients.
•
Hepatocellular carcinoma can still occur after antiviral therapy in patients with cACLD.
•
Non-invasive tests at follow-up can stratify the risk of hepatocellular carcinoma.
•
Portal hypertension-related decompensation is rare after treatment in patients with cACLD.
Background & Aims
It is important to know which patients with hepatitis C are likely to develop liver-related complications after achieving a sustained virological response (SVR) to direct-acting antiviral (DAA) therapy. We aimed to describe the incidence of liver-related events in a population of patients with HCV-associated compensated advanced chronic liver disease (cACLD) who achieved SVR and to identify non-invasive parameters that predict the occurrence of liver-related events.
Methods
This 2-center prospective study included 572 patients with cACLD who had been treated with DAAs and had achieved SVR. Patients had liver stiffness measurement (LSM) ≥10 kPa at baseline and had never decompensated (Child-Pugh class A). Laboratory work-up and LSM were performed at baseline and at 1 year of follow-up.
Results
The median follow-up was 2.8 years during which 32 patients (5.6%) presented with a liver-related event. The incidence rate (IR) of portal hypertension-related decompensation was 0.34/100 patient-years. These patients all had baseline LSM >20 kPa, and LSM did not improve during follow-up in 4 out of 5 of them. Hepatocellular carcinoma (HCC) occurred in 25 patients (IR 1.5/100 patient-years). Albumin levels at follow-up (hazard ratio [HR] 0.08; 95% CI 0.02–0.25) and LSM <10 kPa at follow-up (HR 0.33; 95% CI 0.11–0.96) were independently associated with the risk of HCC. Combining both predictors identified 2 groups with differing risk of HCC occurrence: those with LSM ≥20 kPa at follow-up or those with LSM between 10–20 kPa and albumin levels <4.4 g/dl were at the highest risk (IR ≥1.9/100 patient-years). Visual nomograms predicting HCC risk based on LSM and albumin at 1 year of follow-up were constructed.
Conclusion
In patients with HCV-related cACLD who have achieved SVR with DAAs, HCC is the most frequent liver-related event. Both albumin levels and LSM are useful for stratifying patients based on their risk of developing HCC during follow-up.
Lay summary
New oral antivirals can cure chronic hepatitis C infection, however patients with advanced chronic liver disease are still at risk of presenting with liver-related complications. The most frequent complication after oral antiviral therapy in asymptomatic patients with advanced chronic liver disease was liver cancer. The use of simple parameters such liver stiffness and albumin levels after treatment can help to identify patients at higher or lower risk of liver cancer.
Prediction models or scores for hepatocellular carcinoma (HCC) risk stratification are needed to guide individualized management of patients with HCV-related cirrhosis, who are increasingly being treated with highly effective and safe direct-acting antivirals (DAAs).1,2
We read with great interest the article by M. Pons and colleagues in which they estimated the incidence of liver-related events and predictive ability of non-invasive markers of liver fibrosis after successful antiviral therapy for HCV infection.1 In this prospective 2 center study from Spain, including patients with HCV-related compensated advanced chronic liver disease (liver stiffness (LS) ≥10 kPa), hepatocellular carcinoma (HCC) developed in 4.4% (25/572) of cases within 2.8 years of the end of treatment (EOT).
In the era of direct-acting antivirals (DAA) and substantially improved sustained virological response (SVR) rates, chronic hepatitis C (CHC) patients can be treated with different DAA regimens even if suffering from decompensated cirrhosis. Current guidelines recommend surveying CHC patients with elevated liver stiffness measurements (LSM) ≥10 kPa by transient elastography (TE), indicating possible stage 3 or 4 liver fibrosis, for hepatocellular carcinoma (HCC) with an abdominal ultrasound every 6 months.
We really appreciate Charette et al.'s1 interest in our study.2 We would like to clarify, that our initial aim was to identify a subpopulation among compensated advanced chronic liver disease (cACLD) patients with cured HCV who could avoid hepatocellular carcinoma (HCC) surveillance because of a very low risk of developing HCC during follow up; unfortunately, we failed. All such patients need continued HCC screening.
We thank Nabatchikova et al.1 and Shiba et al.2 for their interest and comments on our work.3 First, we would like to explain that the main reason for predicting an event (hepatocellular carcinoma [HCC] in this case) is to guide future action (for example HCC screening). In our study, our initial aim was to identify a subpopulation of patients with compensated advanced chronic liver disease (cACLD), in whom HCV had been cured, who could avoid HCC surveillance; unfortunately, we failed. All HCV-cured patients with cACLD need continued HCC screening.
Direct-acting antivirals (DAAs) have become the new standard of care for patients with chronic hepatitis C virus (HCV) infection, demonstrating high efficacy, with most patients achieving a sustained virological response (SVR) regardless of HCV genotype. Due to their good safety profile, any patient in any stage of chronic liver disease (from mild fibrosis to decompensated cirrhosis) can be treated with DAAs.
Therefore, it is important to know which patients will be prone to developing liver-related complications, such as hepatocellular carcinoma (HCC) or liver decompensation, requiring lifelong follow-up and which patients can be safely discharged from follow-up.
Several studies have demonstrated that in patients with cirrhosis who have achieved SVR after DAA therapy there is a decrease in liver-related events due to an improvement in liver function and portal hypertension, and a decrease in the incidence of de novo HCC, accompanied by an overall increase in survival rates.
AGA clinical practice update on interaction between oral direct-acting antivirals for chronic hepatitis C infection and hepatocellular carcinoma: expert review.
These studies have also demonstrated that patients with cirrhosis are at higher risk of complications than those without cirrhosis. However, the variability in the methods used to select and define patients with cirrhosis in different published studies and the heterogeneity of patients included, mixing patients with compensated cirrhosis and decompensated cirrhosis, makes it hard to find predictors to identify high-risk populations and validate the results.
It is also important to note that currently most of the asymptomatic patients with advanced fibrosis or cirrhosis are diagnosed by elastographic methods such transient elastography and not by liver biopsy. So in many of these cases, the exact stage of the disease is suspected, but unconfirmed. For this reason, the Baveno VI consensus defined patients with liver stiffness measurement (LSM) ≥10 kPa and no prior decompensation as having “compensated advanced chronic liver disease” (cACLD).
Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension.
This patient population is of particular interest because, although they have never decompensated, approximately 80–90% of them have portal hypertension and up to 50–60% have clinically significant portal hypertension (CSPH).
It would probably be easier to identify low-risk groups for liver-related events in this cACLD population, where some patients have an initial degree of advanced liver disease and higher odds of improving with therapy, than to identify them in mixed cirrhotic populations (Child-Pugh class A-B) affected by more advanced disease (with a higher proportion of CSPH and a greater risk of complications), in whom mechanisms of hepatocarcinogenesis might already have been initiated.
The aim of the present study was to describe the incidence of liver-related events in a population of patients with HCV-associated cACLD who achieved SVR after DAA therapy, and to identify non-invasive parameters (serological and elastographic) to predict the occurrence of liver-related events.
Patients and methods
This is a prospective cohort study from 2 tertiary hospitals (Hospital Universitari Vall d’Hebron and Hospital Clinic) from Barcelona, Spain. All adult patients (≥18 years old) who started DAA therapy for HCV infection between January 1st, 2015 and March 31st, 2016 were assessed for participation in the study. The inclusion criteria were: i) suspected cACLD defined by LSM ≥10 kPa and no prior decompensation (ascites, variceal bleeding, hepatic encephalopathy or jaundice) according to Baveno VI definition;
Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension.
ii) confirmed SVR 12 weeks after finishing therapy; and iii) Child-Pugh class A. Patients were excluded when i) LSM was not available before starting therapy; ii) they had history of prior HCC or developed HCC before confirming SVR; iii) they had prior liver transplant or iv) they had concomitant coinfection with HBV and/or HIV. All consecutive patients who agreed to participate in the study were included. The study was conducted in accordance with the Declaration of Helsinki. The ethical review boards of each participating center approved the protocol. All patients gave written informed consent before inclusion.
Study procedures
Data on demographics (age, gender, body mass index), HCV genotype, type of DAA therapy and laboratory parameters before starting therapy, such as platelet count, renal function, aminotransferases, international normalized ratio, albumin, bilirubin, Child-Pugh and model for end-stage liver disease scores were collected to define baseline characteristics. Data on abdominal ultrasound performed within 6 months before starting therapy was also collected. SVR was defined as undetectable HCV RNA at 12 weeks after finishing treatment using the Cobas AmpliPrep/Cobas TaqMan (Roche Molecular Diagnostics, Pleasanton, CA; lower limit of detection, 15 IU/ml). Once SVR was confirmed, patients underwent abdominal ultrasound and laboratory work-up every 6 months, as per standard clinical practice, and LSM at 12 months after finishing treatment.
Liver stiffness measurement
LSM was performed using transient elastography (Fibroscan 502 Touch, Echosens, Paris, France) at baseline (within 6 months before starting treatment) and at 1 year after finishing treatment, in a fasting state, performed by a single experienced operator in each center and according to the usual standard procedure. Quality criteria used in both centers were: at least 10 valid measurements and an interquartile to median ratio ≤30%. Only valid assessments were considered for the analysis. During follow-up, improvement in LSM was defined as a decrease of ≥20% from basal LSM. This criterion was used because values above and below 15% were considered as a normal variability of the procedure (defined per the interquartile to median ratio of 30%).
Liver-related events and follow-up
During follow-up any liver-related event was registered. Liver-related events were defined as occurrence of ascites, variceal bleeding, hepatorenal syndrome, hepatic encephalopathy or HCC. Follow-up started when DAA therapy was finished, however only liver-related events occurring after SVR was confirmed were considered. Mortality was also registered as liver-related or non-liver-related. Patients were followed until the first occurrence of liver-related events, death or until the end of December 2018, whichever event occurred first. In patients who were lost during follow-up, the last day they were known to be alive was registered.
Statistical analysis
For the descriptive analysis, quantitative variables were expressed as mean (SD) and qualitative variables as absolute frequency and percentage. Comparisons between groups for continuous variables were performed using Student's t test and for categorical variables using Chi-square test or Fisher's test, when appropriate. Paired data (LSM at baseline and at follow-up) were analyzed using pair t test.
Due to different mechanisms of occurrence and different factors that could influence its occurrence, liver-related events were described separately according to those related to portal hypertension (ascites, variceal bleeding, hepatorenal syndrome, hepatic encephalopathy) and HCC events. The Kaplan-Meier method was used to estimate the cumulative incidence of liver-related events in the cohort. Comparison of the survival curves between groups was performed using the log-rank (Mantel-Cox) test. Data was censored when any of the following: liver-related events, death, lost or end of follow-up (until 31st December, 2018), occurred first. Univariate and multivariable Cox regression analyses were performed to identify predictors of liver-related events. Variables found to be statistically significant in univariate analyses (defined as p <0.1) were included in the multivariate analysis by forward stepwise approach. Significant variables (p <0.05) were kept in the final Cox model to assess the hazard of liver-related events. The results are expressed as adjusted hazard ratio (HR) and their 95% CIs. Significance was considered as 2-sided p values <0.05. A nomogram was constructed to predict the risk of HCC during follow-up.
Statistical analyses were performed using STATA 13.1 statistical software (StataCorp, College Station, TX, US).
Results
Study population
A total of 1,563 patients in both centers who started DAA therapy between January 1st, 2015 and March 31st, 2016 were evaluated for eligibility. SVR was confirmed in 1,518 patients (97.1%) and finally, a total of 572 patients with cACLD were included in the study (Fig. 1). The baseline characteristics of the patients included are described in Table 1. The mean age of the patients included was 63.7 years and 49.3% were male. Most patients had HCV genotype 1 (85.8%) and 78.1% patients were treated with DAAs for 12 weeks. Some differences were detected in both cohorts. Patients from Hospital Clínic had lower platelet counts, higher bilirubin, alanine aminotransferase and albumin levels and higher LSM at baseline than in the Vall d’Hebron cohort.
Fig. 1Study flowchart from both Hospital Vall d’Hebron (left) and Hospital Clinic (right) cohorts. DAA, direct-acting antiviral; HCC, hepatocellular carcinoma; LSM, liver stiffness measurement; SVR, sustained virological response.
Table 1Baseline characteristics of the patients included in the study.
Characteristics
Total cohort n = 572
Vall d’Hebron Hospital n = 290
Hospital Clinic n = 282
p value
Age, years
63.7 (11.1)
63.7 (12.1)
63.8 (10.1)
0.895
Body mass index, kg/m2
26.7 (4.3)
27 (4.3)
26.4 (4.3)
0.105
Male, n (%)
282 (49.3)
144 (49.7)
138 (48.9)
0.864
Diabetes, n (%)
125 (21.9)
61 (21)
64 (22.7)
0.632
Arterial hypertension, n (%)
262 (45.8)
135 (46.6)
127 (45)
0.717
Dyslipidemia, n (%)
49 (8.7)
24 (8.3)
25 (9)
0.873
Obesity (BMI ≥30 kg/m2), n (%)
98 (17.1)
51 (17.6)
47 (16.7)
0.771
Alcohol abuse, n (%)
27 (4.7)
12 (4.1)
15 (5.3)
0.641
HCV genotype, n (%)
0.216
1
490 (85.8)
243 (83.8)
247 (87.9)
2
10 (1.8)
5 (1.7)
5 (1.8)
3
39 (6.8)
20 (6.9)
19 (6.8)
4
32 (5.6)
22 (7.6)
10 (3.6)
DAA therapy, n (%)
<0.001
SOF + simeprevir
138 (24.1)
96 (33.1)
42 (15)
SOF + ledipasvir
226 (39.5)
140 (48.3)
86 (30.5)
SOF + daclatasvir
38 (6.6)
22 (7.6)
16 (5.7)
PTV/r/O/D
130 (22.7)
21 (7.2)
109 (38.7)
Other
40 (7)
11 (3.8)
29 (10.3)
Ribavirin, n (%)
389 (68.1)
182 (63)
207 (73.4)
0.007
Treatment duration, n (%)
0.967
12 weeks
447 (78.1)
223 (76.9)
224 (79.4)
24 weeks
125 (21.9)
67 (23.1)
58 (20.6)
Varices*, n (%)
<0.001
No
168 (57.7)
83 (72.8)
85 (48)
Grade I
89 (30.6)
23 (20.2)
66 (37.3)
Grade II
8 (2.8)
7 (6.1)
1 (0.6)
Grade III
26 (8.9)
1 (0.9)
25 (14.1)
Platelet count, x109/L
139.9 (62.1)
155.1 (65.9)
124.3 (53.8)
<0.001
INR
1.08 (0.13)
1.03 (0.09)
1.14 (0.13)
<0.001
Bilirubin, mg/dl
0.9 (0.4)
0.9 (0.4)
1.0 (0.4)
<0.001
ALT, IU/L
103 (67)
93 (65.9)
113 (67.4)
0.001
Albumin, g/dl
4.0 (0.4)
4.0 (0.4)
4.1 (0.4)
<0.001
FIB-4 score
5.6 (4.4)
5.1 (4.4)
6.2 (4.4)
0.002
APRI score
2.6 (2.4)
2.2 (2.4)
2.9 (2.4)
<0.001
Child Pugh score, n (%)
0.447
5
521 (91.1)
261 (90)
260 (92.2)
6
51 (8.9)
29 (10)
22 (7.8)
Baseline LSM, kPa
20.2 (10.4)
18.9 (10.1)
21.4 (10.6)
0.004
LSM ≥20 kPa, n (%)
212 (37.1)
90 (31)
122 (43.3)
0.002
*Data on previous upper gastrointestinal endoscopy was obtained from 114 patients in Vall d’Hebron Hospital and 177 in Hospital Clinic. Continuous variables are expressed as mean (standard deviation). P value obtained using Student’s t test or Chi-square test/Fisher’s test when appropriate. APRI, aspartate aminotransferase to platelet ratio index; ALT, alanine aminotransferase; BMI, body mass index; DAA, direct-acting antivirals; HCV, hepatitis C virus; INR, international normalized ratio; LSM, liver stiffness measurement; PTV/r/O/D, paritaprevir/ritonavir/ombitasvir/dasabuvir; SOF, sofosbuvir.
Alcohol abuse: ≥30 g/day for women and ≥40 g/day for men.
At baseline, mean LSM was 20.2 kPa (SD 10.4 kPa). LSM was repeated in 554 patients after 1 year of follow-up, improving to a mean of 13.9 kPa (SD 9.2 kPa) (p <0.001), with a mean decrease from baseline of 28.9%. LSM improved (defined as decrease from baseline ≥20%) in 391 patients (70.6%). Fig. 2 depicts distribution of LSM at baseline and at follow-up according to LSM value in patients with both measurements; remarkably, almost 40% of the patients achieved LSM <10 kPa at follow-up.
Fig. 2Distribution of LSM in the whole cohort. (A) Proportion of patients in 3 categories according to LSM at baseline and at first year of follow-up (n = 554). (B) Proportion of patients in different ranges of LSM. LSM, liver stiffness measurement.
The median follow-up was 2.9 years (range 0.3–3.8 years) during which 17 patients (3%) were lost to follow-up, 5 patients (0.9%) died from non-liver-related causes and 32 patients (5.6%) presented liver-related events. No patients died from liver-related causes.
The most frequent liver-related event was HCC, which occurred in 25 patients (4.4%), at an incidence rate of 1.5/100 patient-years. The median time to HCC occurrence was 1 year (range 0.4–3.3 years). There were 5 patients (0.9%) who presented with portal hypertension-related decompensation (2 patients suffered from variceal bleeding and 3 patients had ascites). The incidence rate for these decompensations was 0.31/100 patient-years. Two additional patients developed a cholangiocarcinoma (0.3%) during follow-up (incidence rate 0.1/100 patient-years).
Factors related to portal hypertension-related decompensation
Due to the low incidence of events (5 patients; 0.9%), it was not possible to find predictors of occurrence of portal hypertension-related events. All 5 of these patients had baseline LSM >20 kPa. During follow-up, LSM only improved (decrease ≥20%) in 1 of the 5 patients. Four patients had baseline platelet count below 100x109/L, whereas the patient whose LSM improved had a baseline platelet count of 160 × 109/L. Characteristics of these patients are described in Table S1.
Based on LSM, patients who had baseline LSM ≥20 kPa and a lack of LSM improvement during follow-up were at an increased risk of developing liver decompensation (HR 39.7; 95% CI 4.4–355.4) (Fig. S1). The estimated incidence in this high-risk group was 2.7/100 patient-years vs. 0.07/100 patient-years in the low-risk group (p <0.001).
Predictors of HCC occurrence
As shown in Table 2, by univariate analysis, patients who developed HCC during follow-up were older, had lower baseline and follow-up albumin levels and less frequently had LSM <10 kPa at the first year of follow-up. Liver stiffness at follow-up was higher in patients who presented with HCC, but these differences were not statistically significant. By multivariate analysis, at baseline, only albumin levels (HR 0.29; 95% CI 0.11–0.76; p = 0.012) were independently associated with the risk of presenting with HCC during follow-up. At follow-up, albumin levels (HR 0.08; 95% CI 0.02–0.25; p <0.001) and LSM <10 kPa at follow-up (HR 0.33; 95% CI 0.11–0.96; p = 0.042) were the only predictors associated with the risk of presenting with HCC during follow-up.
Table 2Factors associated with risk of presenting HCC occurrence.
Characteristics
No HCC n = 547
HCC n = 25
p value
Multivariate analysis HR (95% CI; p value)
At baseline
Age, years
63.5 (11.2)
69.2 (7.2)
0.011
1.04 (1.0–1.1; 0.064)
Body mass index, kg/m2
26.7 (4.2)
27.4 (5.2)
0.438
–
Male, n (%)
269 (49.2)
13 (52)
0.783
–
Arterial hypertension, n (%)
249 (45.5)
13 (52)
0.526
–
Diabetes, n (%)
117 (21.4)
8 (32)
0.210
–
Dyslipidemia, n (%)
46 (8.4)
3 (12)
0.226
–
Obesity (BMI ≥30 kg/m2), n (%)
92 (16.8)
6 (24)
0.323
–
Alcohol abuse, n (%)
26 (4.8)
1 (4)
0.435
–
Platelet count, ×109/L
139.3 (61.2)
152.4 (80.8)
0.305
–
Platelet <150×109/L, n (%)
347 (63.4)
15 (60)
0.728
–
Bilirubin, mg/dl
0.9 (0.4)
1.0 (0.3)
0.388
–
ALT, IU/L
104 (67.9)
86 (46.1)
0.202
–
Albumin, g/dl
4.1 (0.4)
3.9 (0.5)
0.003
0.29 (0.11–0.76; 0.012)
FIB4
5.6 (4.5)
5.9 (4.1)
0.745
–
APRI
2.6 (2.5)
2.2 (1.5)
0.452
–
MELD = 6, n (%)
132 (24.1)
5 (20)
0.705
–
Baseline LSM, kPa
20.0 (10.4)
22.5 (9.4)
0.245
–
Baseline LSM <20 kPa, n (%)
348 (63.6)
12 (48)
0.114
–
At follow-up
Body mass index, kg/m2
27.3 (4.2)
27.9 (5.5)
0.655
–
Obesity (BMI ≥30 kg/m2), n (%)
96 (17.6)
6 (24)
0.386
–
Platelet count, ×109/L
172 (70.4)
157.4 (64.6)
0.470
–
Bilirubin, mg/dl
0.7 (0.4)
0.8 (0.3)
0.908
–
Albumin, g/dl
4.4 (0.3)
4.1 (0.4)
<0.001
0.08 (0.02–0.25; <0.001)
Follow-up liver stiffness, kPa
13.8 (9.2)
17.1 (8.3)
0.089
–
Delta LSM, %
−29.2 (34.2)
−22.4 (33.3)
0.355
–
LSM improvement*, n (%)
378 (71.2)
13 (56.5)
0.131
–
Follow-up LSM <10 kPa, n (%)
216 (40.7)
4 (17.4)
0.025
0.33 (0.11–0.96; 0.042)
Values in bold indicate statistical significance. Continuous variables expressed as mean (SD). p value obtained using Cox regression analysis. APRI, aspartate aminotransferase to platelet ratio index; ALT, alanine-aminotransferase; BMI, body mass index; HCC, hepatocellular carcinoma; HR, hazard ratio; LSM, liver stiffness measurement; MELD, model of end-stage liver disease.
Alcohol abuse: ≥30 g/day for women and ≥40 g/day for men. *LSM improvement: Decrease in LSM ≥20% at follow-up, compared to baseline.
In order to construct a simple model to predict the risk of HCC, survival curves among different subgroups based on multivariate analysis were evaluated. We categorized baseline LSM based on established cut-offs (<15 kPa, 15–20 kPa, ≥20 kPa) to compare risk groups.
Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension.
Thus, LSM >15 kPa was highly suggestive of cACLD and LSM ≥20 kPa indicated a very high probability of CSPH. As detected in multivariate analysis, no differences were observed in HCC occurrence according to baseline LSM (Fig. S2A) (p = 0.287). Moreover, there were no differences in HCC occurrence in patients according to LSM improvement, neither when improvement was defined as a decrease in ≥20% (p = 0.117) nor when defined as recently published (decrease in ≥30%) (p = 0.274) (Fig. S2B, C).
As previously shown in Table 2, HCC risk was different in patients according to baseline albumin. When selecting the cut-off of 4 g/dl based on the mean value, patients with albumin <4 g/dl at baseline had an increased risk of HCC occurrence compared to those with baseline albumin ≥4 g/dl (HR 3.27; 95% CI 1.45–7.36; p = 0.004) (Fig. 3A). Meanwhile, the risk of developing HCC differed according to LSM during follow-up, observing a linear trend, that is, the higher the LSM at follow-up the higher the risk of HCC (Fig. 3B). Patients with an LSM of between 10–20 kPa at follow-up had an increased, but not significant, risk of HCC compared to those with LSM <10 kPa at follow-up (HR 2.48; 95% CI 0.79–7.80; p = 0.120), while the HR for HCC was 4.53 (95% CI 1.36–15.08, p = 0.014) for patients with LSM ≥20 kPa at follow-up compared to those with LSM <10 kPa (Fig. 3B). HCC risk was also different according to albumin levels during follow-up. Patients with albumin <4.4 g/dl (based on the mean value) were at a high-risk of developing HCC (HR 2.36; 95% CI 1.02–5.47; p = 0.046) (Fig. 3C).
Fig. 3Cumulative incidence of HCC occurrence according to different subgroups. (A) Albumin levels at baseline; (B) Follow-up LSM; (C) Follow-up albumin levels. HCC, hepatocellular carcinoma; LSM, liver stiffness measurement. (This figure appears in colour on the web.)
We combined both follow-up albumin and LSM to identify risk groups according to predictors gathered at the same time point (1 year after finishing treatment). Table 3 shows the incidence rate of HCC according to different combinations of albumin and LSM at follow-up. Incidence rates of HCC were <1/100 patient-years in patients with LSM <10 kPa at follow-up and in those with LSM between 10–20 kPa and high albumin levels (≥4.4 g/dl at baseline). On the contrary, patients with LSM ≥20 kPa at follow-up and those with LSM 10–20 kPa and low albumin levels (<4.4 g/dl) have incidence rates of HCC ≥1.9/100 patient-years, indicating that albumin levels only impacted on HCC risk in patients with LSM between 10–20 kPa. Therefore, we identified 2 main risk groups, a low-risk group and a high-risk group for HCC, with different estimates of HCC incidence (Fig. 4). A Cox regression model was constructed using these variables (albumin <4.4/≥4.4 and LSM at follow-up <10 kPa/10–20 kPa/≥20 kPa) with a good predictive power (Harrell’s C index = 0.73) (Fig. S3). In Fig. 5, Fig. 2 nomograms to predict the probability of HCC occurrence according to albumin levels and LSM at follow-up are shown.
Table 3Incidence rates of HCC in the cohort according to different subgroups (both albumin and LSM at follow-up).
Fig. 4Cumulative incidence of HCC according to low-risk group or high-risk group. Low-risk group: follow-up LSM <10 kPa or follow-up LSM 10–20 kPa + follow-up albumin ≥4.4 g/dl. High-risk group: follow-up LSM ≥20 kPa or follow-up LSM 10–20 kPa + follow-up albumin <4.4 g/dl. HCC, hepatocellular carcinoma; LSM, liver stiffness measurement. (This figure appears in colour on the web.)
The main findings of the present study, in a cohort of patients with HCV-associated cACLD who achieved SVR after DAA therapy, were that the incidence of liver-related events was relatively low, with de novo HCC being the most frequent complication, and that using simple non-invasive predictors (albumin and LSM at follow-up) we could identify 2 patient groups with a differing risk of HCC during follow-up.
Several studies have demonstrated that in patients who achieve SVR after DAA therapy, the incidence of HCC decreases and that the main predictor of its occurrence, as expected, is the presence of cirrhosis, with patients affected with a more advanced liver disease (such as decompensated cirrhosis) and those with a history of HCC at the highest risk.
Impact of previously cured hepatocellular carcinoma (HCC) on new development of HCC after eradication of hepatitis C infection with non-interferon-based treatments.
HCC was the main complication observed in our study. There were 25 patients (4.4%) who developed HCC, that is 78% of all liver-related events observed in the cohort. The incidence rate of HCC was 1.5/100 patients-year. This incidence was similar to previous reported series which included patients with cirrhosis.
In contrast to what has been previously reported, when we stratified the risk of HCC according to LSM at baseline, we did not find differences in HCC incidence among patients with LSM ≥20 kPa and those with LSM <20 kPa. Two studies reported that patients with high baseline LSM had an increased risk of HCC.
Factors associated with increased risk of de novo or recurrent hepatocellular carcinoma in patients with cirrhosis treated with direct-acting antivirals for HCV infection.
Factors associated with increased risk of de novo or recurrent hepatocellular carcinoma in patients with cirrhosis treated with direct-acting antivirals for HCV infection.
In our cohort, 212 patients had baseline LSM ≥20 kPa, of whom 13 (6.1%) presented with HCC, while HCC occurred in 12 patients with baseline LSM <20 kPa (3.3%); these differences were not statistically significant. One explanation for these differences between our study and previously published studies is that in both previous studies Child-Pugh B patients were included. Although the distribution of LSM was not described according to Child-Pugh class, those patients with worse liver function also had higher LSM, which could lead to an overestimation of the effect of baseline LSM. In our cohort, including only patients with Child-Pugh class A and no prior history of HCC, baseline LSM did not predict the occurrence of HCC.
A retrospective study from Ravaioli, et al. including 139 patients with cirrhosis (11.5% Child-Pugh B and 13.7% with previous HCC) reported that a decrease of 30% in LSM was one of the predictors that was inversely associated with the risk of presenting with HCC.
In our cohort, LSM improvement (defined as LSM decrease ≥20%) was not associated with the risk of HCC. LSM improved in 13 patients (52%) who developed HCC and in most of the patients who did not develop HCC (71.2%). This probably reflects that most of the early dynamic changes in LSM are related to improvement in liver inflammation and they do not accurately reflect improvement in fibrosis. It is important to note that these high rates of LSM improvement do not translate into a decreased risk of HCC and for this reason it is important to continue HCC surveillance.
The main predictors related to HCC risk in our study were albumin and follow-up LSM. Although most of the patients had normal albumin levels, the risk of HCC was reduced by 58% in patients with albumin ≥4.4 g/dl at follow-up. Based on both predictors (albumin and LSM at follow-up), we could identify 2 main risk groups: i) patients with LSM at follow-up <10 kPa or patients with LSM between 10 and 20 kPa with albumin ≥4.4 g/dl at follow-up had a lower incidence of HCC (0.6/100 patients-year) compared to ii) patients with LSM ≥20 kPa at follow-up or those with LSM between 10–20 kPa but albumin <4.4 g/dl at follow-up (2.9/100 patients-year). It has to be highlighted that 220 patients achieved LSM <10 kPa during follow-up and of them, 4 patients (1.8%) developed HCC. Although the risk is much lower than in patients with the highest LSM at follow-up, the recommendation, as in recently published guidelines, is to continue HCC surveillance after SVR in patients with pre-treatment LSM ≥10 kPa until information about longer follow-up is available.
AGA clinical practice update on interaction between oral direct-acting antivirals for chronic hepatitis C infection and hepatocellular carcinoma: expert review.
However, a recent study reported that surveillance for patients with advanced fibrosis (i.e. F3) who achieved SVR is not cost-effective due to the low incidence of HCC, questioning the need for continued surveillance in some specific subgroups of patients.
Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
Metabolic syndrome has been associated as a risk factor of HCC occurrence in patients who have achieved SVR, as was observed in the French CirVir cohort.
One explanation could be the lack of improvement of fibrosis in these patients and an increase of non-alcoholic steatohepatitis prevalence in patients with these features. However, we could not observe an association between some factors of metabolic syndrome such as body mass index, diabetes or dyslipidemia and the occurrence of HCC in our cohort. The main differences between our cohort and the Nahon et al. cohort are the number of patients included (>1,000 in the Nahon et al. study) and the time of follow-up (median 58.2 months in the Nahon et al. study). Probably, a larger follow-up will help to better define the association between metabolic syndrome and HCC in patients with cACLD.
The incidence of portal hypertension-related complications was very low in this cohort, since only 5 patients (0.9%) presented with liver decompensation (3 ascites and 2 variceal bleeding). This is the first prospective study in patients with cACLD that reports portal hypertension complications after achieving SVR. In previous studies it has been demonstrated that SVR after DAA therapy reduces the incidence of portal hypertension-related complications.
However, contrary to the present study, most of the series that have evaluated the incidence of portal hypertension-related complications after DAA therapy have included patients with compensated and decompensated cirrhosis and some of the predictors that had been related to decompensation, such as previous decompensation, cannot be applied in this cohort with cACLD.
Durability of virologic response, risk of de novo hepatocellular carcinoma, liver function and stiffness 2 years after treatment with ombitasvir/paritaprevir/ritonavir±dasabuvir±ribavirin in the AMBER, real-world experience study.
Although we could not find predictors of decompensation due to the low incidence rates, all 5 decompensated patients had baseline LSM ≥20 kPa and in 4 out of 5 patients LSM did not improve after SVR, remaining >20 kPa. These 4 patients also have a platelet count below 100x109/L. This suggests that these patients had more advanced liver disease and SVR did not reduce the risk of decompensating. Although values of LSM have not yet been validated in non-viremic patients, LSM >20 kPa after SVR seems to have high specificity in ruling in the presence of CSPH and our results support this evidence.
Portal pressure and liver stiffness measurements in the prediction of fibrosis regression after sustained virological response in recurrent hepatitis C.
There was 1 patient, however, who developed variceal bleeding during follow-up and LSM improved from 23.1 kPa to 16.3 kPa during follow-up. Indeed, Lens et al. reported that in patients with LSM <13.6 kPa at 6 months of follow-up, CSPH was still present in 43% of patients.
Our study has some limitations. LSM was performed after 1 year of finishing treatment and therefore we do not have information about dynamics before and beyond that. The earliest a predictor can be determined, the more useful it is. Considering that most of the LSM decrease after SVR occurs early (at the end of therapy)
and that LSM values do not change much afterwards, it is plausible to assume that the same predictions about liver-related events presented here can be done using LSM values obtained right after finishing therapy. Another limitation is that the current results can only be applied to patients in whom an LSM can be obtained and is reliable. Also, the validation of the predictors using an external cohort would definitely strengthen the results. Instead, we can confirm that the predictors for HCC of our study were valid and significant for the 2 separate cohorts. Fig. S4 depicts the cumulative incidence of HCC in each hospital according to low-risk and high-risk group definitions. Despite these limitations, the strength of our study relies on being the first prospective bicentric study reporting outcomes after SVR of a large cohort of patients with cACLD, Child-Pugh A, with a follow-up of up to 3.5 years, providing useful information about the prognosis and simple prediction of events in asymptomatic patients with advanced liver disease.
In conclusion, the present study demonstrates that in patients with cACLD who have achieved SVR with DAA, HCC occurrence is the most frequent liver-related complication. Both albumin levels and LSM during follow-up can help to identify patients at the highest risk of presenting with HCC. Although patients with LSM <10 kPa at follow-up have a low risk of presenting HCC, a zero-risk subpopulation cannot be found, thus, HCC surveillance needs to be considered for all patients classified as having cACLD before therapy. However, with the information provided by our study, patients and clinicians will possess the right information about the expected risk of HCC on an individual basis and more importantly, it will serve to stress the importance of continuing HCC screening (especially in the high-risk groups) and maintaining adherence to these programs by patients.
Abbreviations
APRI, aspartate aminotransferase to platelet ratio index; ALT, alanine aminotransferase; BMI, body mass index; cACLD, compensated advanced chronic liver disease; CSPH, clinically significant portal hypertension; DAA, direct-acting antiviral; HCC, hepatocellular carcinoma; HR, hazard ratio; INR, international normalized ratio; IR, Incidence rate; LSM, liver stiffness measurement; PTV/r/O/D, paritaprevir/ritonavir/ombitasvir/dasabuvir; SOF, sofosbuvir; SVR, sustained virological response.
Financial support
MP is a recipient of a Río Hortega grant from Instituto de Salud Carlos III, Spain. SRT received a grant from Instituto de Salud Carlos III, Rio Hortega program CM17/00015, and an Initiation research grant from the Catalonian Society of Digestology and the Emili Lentag end-of-residency prize from Hospital Clínic de Barcelona. SL was supported by a grant from Hospital Clinic. XF received support by Secretaria d'Universitats i Recerca del Departament d'Economia i Coneixement (grant 2017 SGR 1753) and CERCA Programme/Generalitat de Catalunya. JG is a recipient of a Research Intensification grant from the Instituto de Salud Carlos III. The work was partially funded by grants PI15/00066, PI18/00079, PI18/00961 and PI18/00947 from Instituto de Salud Carlos III and co-funded by European Union (ERDF/ESF, “Investing in your future” – Una manera de hacer Europa). CIBERehd is supported by Instituto de Salud Carlos III.
Conflicts of interest
ZM is an advisor for Gilead and Abbvie. VV is a speaker for Intercept and an advisor for Promethera. SL has received speaker/advisory fees from Abbvie, Gilead, Janssen and MSD. MB is an advisor and speaker for Gilead, MSD and Abbvie. XF has served as advisor for AbbVie and Gilead. BM has received lecture fees from Gilead and Bayer and is an advisor for Bayer.
Please refer to the accompanying ICMJE disclosure forms for further details.
Authors’ contributions
Study concept and design: MP, JG. Patient enrollment and acquisition of data: MP, SRT, JIE, ZM, VV, SL, MB, SA, XF, BM, JG. Analysis and interpretation of data: MP, BM, JG. Drafting of the manuscript: MP, JG. Critical revision of the manuscript: SRT, JIE, ZM, VV, SL, MB, SA, XF, BM. All the authors approved the final draft which is being submitted.
Supplementary data
The following are the Supplementary data to this article:
AGA clinical practice update on interaction between oral direct-acting antivirals for chronic hepatitis C infection and hepatocellular carcinoma: expert review.
Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension.
Impact of previously cured hepatocellular carcinoma (HCC) on new development of HCC after eradication of hepatitis C infection with non-interferon-based treatments.
Factors associated with increased risk of de novo or recurrent hepatocellular carcinoma in patients with cirrhosis treated with direct-acting antivirals for HCV infection.
AGA clinical practice update on interaction between oral direct-acting antivirals for chronic hepatitis C infection and hepatocellular carcinoma: expert review.
Cost effectiveness of hepatocellular carcinoma surveillance after a sustained virologic response to therapy in patients with hepatitis C virus infection and advanced fibrosis.
Durability of virologic response, risk of de novo hepatocellular carcinoma, liver function and stiffness 2 years after treatment with ombitasvir/paritaprevir/ritonavir±dasabuvir±ribavirin in the AMBER, real-world experience study.
Portal pressure and liver stiffness measurements in the prediction of fibrosis regression after sustained virological response in recurrent hepatitis C.
30606-3&id=ga1.jpg){kind=link}
30606-3&id=gr1.jpg){kind=link}
30606-3&id=gr2.jpg){kind=link}
30606-3&id=gr3.jpg){kind=link}
30606-3&id=gr4.jpg){kind=link}
30606-3&id=gr5.jpg){kind=link}