| |
|
DAA therapy and long-term hepatic function in advanced/decompensated cirrhosis: Real-world experience from HCV-TARGET cohort
|
| |
| |
Download the PDF here
JNL of Hepatology April 1 2020
Article in Press
Elizabeth C. Verna, MD, MS, Giuseppe Morelli, MD, Assistant Professor, Norah A. Terrault, MD, Professor, Anna S. Lok, MD, Professor, Joseph K. Lim, MD, Adrian M. Di Bisceglie, MD, Professor, Stefan Zeuzem, MD, Charles S. Landis, MD, PhD, Associate Professor, Paul Kwo, MD, Mohamed Hassan, MD, Professor of Medicine, Michael P. Manns, MD, Professor, Monika Vainorius, MD, Director, Lucy Akushevich, MS, Senior Biostatistician, David R. Nelson, MD, Professor of Medicine, Michael W. Fried M, MD, Professor, K. Rajender Reddy, MD, Professor
In summary, in this real-world cohort of HCV treatment among patients with advanced/decompensated cirrhosis, while SVR rates were high, the improvement in MELD score, bilirubin or albumin after a median follow-up of 4 years were minimal. Thus, while an optimistic position would be that there was no worsening in the severity of liver disease in the majority of patients and few required liver transplant during follow-up, a state of MELD “purgatory” may evolve in some patients and most patients require ongoing close monitoring.
Highlights
• Sustained virologic response was achieved in 90.5% of patients with advanced/decompensated cirrhosis treated with directly acting antiviral therapy
• In long-term follow up (median of 4 years after treatment), overall mean changes in MELD, total bilirubin and albumin were marginal
• A clinically meaningful decrease in MELD of ≥ 3 occurred in 29% and a final MELD score of < 10 was achieved in 25%
• Patients with advanced/decompensated liver disease should continue to be monitored following SVR
Abstract
Background and Aims
Direct-acting antiviral (DAA) HCV therapy is used in decompensated cirrhosis with the expectation of improvement in hepatic function. Little is known about the long-term benefit of successful treatment.
Methods
Patients with advanced/decompensated cirrhosis (MELD ≥10) in HCV-TARGET who initiated NS5A-containing DAA therapy prior to September, 2018, were included. Treatment outcomes and the impact of treatment on short-term and long-term hepatic function were examined.
Results
642 patients were analyzed. The mean age was 60 years, 68% were male. The median baseline MELD was 12 (range 10-39) and 64% had prior decompensation. Among patients with available virologic outcomes, 90.5% achieved SVR12. Twenty-four % achieved a clinically significant decrease in MELD by ≥3 points during short term follow-up (9-26 weeks after the end of treatment). However, in long-term follow up (median of 4 years after treatment), mean changes in MELD (-0.30 points), total bilirubin (+0.23 mg/dl) and albumin (+0.36 g/dl) were marginal. Fifty-one patients died and 22 underwent liver transplant. In long term follow up, a clinically meaningful decrease in MELD of ≥3 occurred in 29% and a final MELD score of <10 was achieved in 25%.
Conclusion
In a large real-world experience of patients with advanced/decompensated HCV cirrhosis treated with DAA, there were only marginal improvements in MELD, total bilirubin, or albumin in long-term follow up (median of 4 years after treatment) after achieving SVR; a clinically meaningful decrease in MELD of ≥3 occurred in 29% and a final MELD score of <10 was achieved in 25%. These patients may remain at high risk of decompensation and must continue to be closely followed.
The multivariable model with LASSO-selected covariates for association with SVR included the presence of ascites (OR 0.58, 95% CI 0.26-1.21, p=0.157), albumin level <3.5 mg/dl (OR 0.41, 95%CI 0.14-1.04, p=0.070), history of HCC (OR 0.40, 95% CI 0.18-0.91, p=0.024), PPI use (OR 0.59, 95% CI 0.31-1.08, p-value =0.099), age<60 (OR 0.58, 95% CI 0.30-1.09, p=0.085), male sex (OR 0.49, 95% CI 0.23-0.98, p=0.049), MELD<16 (OR 0.36, 95% CI 0.11-0.93, p=0.053), and Child-Pugh A (OR 1.84, 95% CI 0.53-7.28, p=0.342)(Supplemental Figure 2).
During short-term follow-up, MELD score decreased in 187 (56%), did not change in 54 (16%) and increased in 92 (28%) patients (Figure 3A). A decline of at least 3 MELD points occurred in 80 (24%) patients and was more common among patients with a baseline MELD ≥16 (41%)(Figure 3A). Short-term MELD change among patients who achieved SVR and who failed DAA treatment were similar (Figure 3B).
In multivariable model with LASSO-selected covariates, male sex (OR 0.51, 95% CI 0.27-0.93, p=0.029), baseline ALT<60 u/l (OR 0.41, 95% CI 0.22-0.75, p=0.004), Child-Pugh score A (OR 0.48, 95% CI 0.21-0.99, p=0.058) and MELD<16 (OR 0.33, 95% CI 0.16-0.69, p=0.003) were negatively associated with MELD improvement of at least 3 points, and age<60 (OR 2.15, 95% CI 1.20-3.93, p=0.012) was predictive of this improvement (Figure 4).
A total of 218 patients in the efficacy population who achieved SVR who were not on dialysis at baseline were evaluable for long-term outcomes. Of these, 158 had MELD scores available at least 36 weeks after the end of treatment. The median (interquartile range) follow-up after the end of treatment was 213 (174-224) weeks. Baseline demographics were similar between patients with and without long-term follow-up, including age (mean 58.5 vs. 59.8, p=0.168), HCC (8.2% vs 11.9%, p=0.290), baseline MELD (mean 13.2 vs. 12.9, p=0.417), and baseline albumin (mean 3.2 vs. 3.3, p=0.257) (Table 2). However, patients with long-term follow up were more likely to have ascites at baseline (55.7% vs. 44.1%, p=0.044) and prior hepatic decompensation events (71.5 % vs. 60.8 %, p=0.050).
The most common liver-related SAEs of interest were the development of hepatic encephalopathy (4.4%), ascites (0.9%), and bacterial peritonitis (0.8%). Overall, 83 patients (12.9%) had new decompensating events either on treatment or up to 30 days after the end of treatment and 41 patients (6.4%) had decompensating events that were classified as SAEs. Of the 83 patients, only 7 patients (8%) had not had any history of decompensating events reported prior to DAA regimen. 94 (69%) of the 137 AEs experienced by 61 (73%) of these 83 patients were exacerbations of previously recorded condition. 15 remaining patients (18%) with prior history of decompensating events had developed a different decompensating event than what was recorded prior to start of DAA treatment.
There was a total of 16 patients with baseline MELD ≥21 treated with elbasvir/grazoprevir. All of these patients had eGFR of <30 mg/dl and 15 of them were on dialysis at baseline. There were no decompensating events characterized as SAEs recorded in this subpopulation.
Survival:
Seven patients died on treatment (coronary artery disease, hepatic encephalopathy, hepatic failure, multi-organ failure, septic shock, and unknown cause in two patients), four of whom had a baseline MELD 10-15, one MELD 16-20 and two MELD ≥ 21. In addition, at least 44 patients died after the end of DAA treatment. Death was attributable to advanced liver disease in 12 patients, while the cause of death in the others was liver unrelated or unknown. Median duration (range) of time for survival follow-up was 15.5 (0–50.8) months. Baseline MELD <16 was predictive of overall survival (log rank, p<0.0001). One-, two-, and three- year survival rates in participants with baseline MELD 10-15 and baseline MELD ≥16 were 96.0%-93.6%-89.7% and 85.9%-82.4%-71.7%, respectively.
Five patients underwent liver transplantation while on treatment. Three achieved SVR, one failed treatment, and virologic data was unavailable in one patient. In addition, 17 patients underwent liver transplant after the end of treatment. Median (range) follow up time for liver transplant events was 14.1 (0–50.8) months. The proportion of patients with baseline MELD 10-15, and baseline MELD ≥16, who underwent liver transplant 1-, 2-, and 3 years after treatment initiation was 2.6%- 5.4%- 9.0% and 1.2%- 2.9%-2.9%, respectively. Median duration (range) of time for HCC diagnosis follow-up was 14.1 (0–50.8) months. Total of 13 patients with baseline MELD 10-15 and 3 patients with baseline MELD > 16 developed HCC.
Discussion:
Chronic Hepatitis C therapy in those with advanced stage of liver disease may achieve improvement in hepatic function, as measured by MELD or CTP scores, although who will achieve this improvement and whether long-term transplant-free survival is a realistic goal of treatment remain uncertain5,6,14, 15, 16, 17, 18, 19. This is a large real-life cohort of patients with HCV treatment in advanced/decompensated liver disease with the longest post-treatment follow up available to date.
The overall SVR12 rate in the efficacy population (90.5%) is similar to those reported in clinical trials and previous real-world experiences.3,4 Remarkably, 97% of the 34 patients with MELD over 21 who had available virologic outcome achieved viral eradication, though this SVR rate decreased to 78.6%, when patient deaths on or after treatment but prior to SVR 12 timepoint were classified as failures. Elbasvir/grazoprevir, which is contraindicated in patients with advanced liver disease, was more commonly used in patients with MELD score of ≥21 because most patients in this high MELD group had advanced kidney disease. Despite high SVR rates, changes in MELD score and hepatic function tests in short- and long-term follow-up in this cohort were marginal. While 56% of patients experienced some improvement in MELD at the time of SVR12, the overall mean short-term change in MELD was less than a point and a clinically meaningful improvement of ≥3 MELD points occurred in only 24% of patients. Predictors of this improvement included female sex, age<60, high baseline ALT (≥60 IU/ml), and baseline MELD ≥16. Our data are in line with clinical trials with a variety of regimens where median MELD improvement, over the short term, was by 2 points in 60% of patients; and MELD score remained unchanged or worsened by a median MELD of 1 in the remaining 40%.3
Unique to this cohort is the long-term follow-up - median >4 years post-treatment. Patients with long-term follow up were not dissimilar from those without, except that ascites and prior decompensating events were more common among patients included in the long-term follow up, thus indicating that those in follow up were as sick if not sicker than those lost to follow up. At a median of 213 weeks of follow-up, there were only minimal overall mean changes in MELD score (decrease 0.30 points), total bilirubin (increase by 0.23 mg/dl) and albumin (increased by 0.36 g/dl). In addition, only 29% of patients experienced a clinically meaningful improvement in MELD by 3 or more points and 25% achieved a MELD score of <10; a threshold we used as a surrogate for reasonably good synthetic function and functional status.
Previous reports had noted improvements in hepatic function, to the degree of de-listing patients on a liver transplant list. Improvements in CPT scores have been noted and predictors of improvement have been described.5, 6, 7, 8 A prior study has evaluated changes in MELD in CPT B and C patients and noted that at 36 months after HCV treatment, a substantial number of patients with MELD ≥15 still remained in that category.20 A few reports have suggested that some patients achieving SVR after DAA had an improvement in portal hypertension below the clinically significant portal hypertension threshold of 10 mm of Hg while MELD score may not reflect such benefit21,22. Unfortunately given the lack of widespread use and availability of hepatic venous pressure measurements in real life scenario, we were limited in generating such data.
While there has been a decrease in listings for liver transplantation with HCV as an etiology in national database, it is unclear if this is because of preventing progression of liver disease and development of HCC in those with compensated cirrhosis or if it is due to improvements in hepatic function in those with decompensated cirrhosis.2,23 Further, listing or de-listing criteria are arbitrary, and center specific and thus difficult to compare across studies and across liver centers. Our data would suggest that there is improvement in hepatic function in some patients but not necessarily to the degree where the risk of death and functional incapacity are mitigated if MELD changes were to be used as a surrogate for these events.
There are limitations to this study inherent to the real-world study design. In particular, the lack of data available to determine change in CTP class over time limits our ability to understand the impact of treatment on non-MELD based indications for liver transplant. We do not have quality of life and liver specific symptom data, particularly pertaining to ascites and hepatic encephalopathy resolution to verify ongoing symptoms of portal hypertension. It is possible that there are patients with elevated MELD scores due to factors other than decompensated liver disease, such as advanced chronic kidney disease. To mitigate this potential misclassification, patients on dialysis were excluded from MELD trajectory analysis.
In summary, in this real-world cohort of HCV treatment among patients with advanced/decompensated cirrhosis, while SVR rates were high, the improvement in MELD score, bilirubin or albumin after a median follow-up of 4 years were minimal. Thus, while an optimistic position would be that there was no worsening in the severity of liver disease in the majority of patients and few required liver transplant during follow-up, a state of MELD “purgatory” may evolve in some patients and most patients require ongoing close monitoring.
Introduction:
The high rates of sustained virologic response (SVR) in patients with HCV-compensated cirrhosis, in both clinical trials and real-world settings, has led to widespread recommendations for treatment in this setting1 and concurrently there has been a decline in new listings for liver transplantation due to HCV-related liver disease.2 However, some controversy exists in the treatment of patients with decompensated cirrhosis due to suboptimal SVR rates, and concerns regarding the impact of SVR on long-term outcomes among liver transplant candidates.3,4
While most studies have reported improvements in MELD score and/or CTP class in a significant proportion of patients3, the impact on the long-term trajectory of liver function, including resolution of portal hypertension, and transplant-free survival remains uncertain. Predictors of short-term resolution in liver dysfunction may include normal body mass index, lack of encephalopathy or ascites at treatment initiation, and normal serum levels of alanine aminotransferase and albumin.5 However, those with severe portal hypertension complications and MELD >20 may be less likely to achieve clinical improvement.6, 7, 8 Concern remains that many patients with life-threatening complications of liver disease may be inadvertently disadvantaged in terms of transplant access by a moderate treatment-related decrease in MELD score and without resolution of cirrhosis complications.3,6,7 Thus, we aimed to determine effectiveness of DAA therapy in a large real-world cohort of patients with advanced/decompensated cirrhosis. In particular, we evaluated the impact of SVR on hepatic function in long-term follow up, at a median of 4 years after completion of treatment.
Methods:
Patients and Study Design:
HCV-TARGET is an international consortium of academic (n=46) and community (n=16) medical centers in the U.S., Germany, Israel and Canada conducting a longitudinal, prospective observational cohort study of real-world administration of DAA therapy (NIH Clinical Trial NCT01474811). All patients were adults ≥18 years, who underwent HCV treatment administered according to local standard of care. Data from enrolled patients undergoing HCV therapy were captured from medical records within a common database utilizing centralized data abstraction.
Patients included in this analysis were those with advanced/decompensated cirrhosis who started HCV treatment between March, 2014 and September, 2018 with an all oral DAA regimen. Patients with prior liver transplantation were excluded.
The independent ethics committee at each participating study center or a central institutional review board approved the protocol. All patients provided written informed consent. All authors had complete access to the study data and reviewed and approved the final manuscript. Study data were collected and managed using REDCap electronic data capture tools9 hosted at the University of North Carolina Chapel Hill. REDCap (Research Electronic Data Capture) is a secure, web-based application.
Advanced/Decompensated Cirrhosis Assessment
Cirrhosis was defined at the time of enrollment by biopsy (METAVIR stage 4 fibrosis) or FibroScan (liver stiffness of ≥12.5 kPa) and/or a combination of clinical, laboratory, histologic, and imaging criteria. Patients with METAVIR stage 3 fibrosis by liver biopsy/Fibroscan were considered to have cirrhosis if they had any of the following: platelet count <140,000/mL; presence of esophageal varices on esophagogastroduodenoscopy; nodular liver, portal hypertension, or ascites by radiologic imaging; or non-invasive serum panels such as FibroSURE (Laboratory Corporation of America, Burlington, NC) consistent with stage 4 fibrosis. In the absence of cirrhosis being confirmed by liver biopsy and/or Fibroscan, cirrhosis was defined as meeting any two of the earlier described non-histologic criteria. Advanced/decompensated cirrhosis was defined as having both cirrhosis and a MELD score ≥10. While there is no MELD cutoff that is known to define decompensated liver disease, this cutoff was chosen as it is associated with at least 6% 3-month mortality and also is low enough to capture patients with complications of portal hypertension that are out of proportion to MELD score elevation10,11.
History of hepatic decompensation was defined as evidence of prior or current diagnosis of ascites, hepatic encephalopathy, spontaneous bacterial peritonitis, variceal hemorrhage, or baseline concomitant medications with a specific indication for ascites, hepatic encephalopathy, spontaneous bacterial peritonitis, or variceal hemorrhage. Baseline Child-Turcotte-Pugh (CTP) score was estimated based on a validated algorithm12, including laboratory values within 90 days of HCV treatment initiation. Clinical records were reviewed up to 6 months prior to HCV treatment initiation for evidence of ascites and encephalopathy, and quantified for severity according to the operational definitions of the algorithm of Kaplan et al. in order to calculate CTP score12.
OUTCOMES:
Efficacy:
Treatment efficacy was measured as sustained virologic response (SVR 12), defined as HCV RNA level below the limit of quantitation or undetected at least 9 weeks after treatment was discontinued (n=577, Figure 1 Efficacy population). Those who were lost to follow up, died or had data missing at the time of this analysis were excluded.
Safety:
All patients who started therapy during the study period and completed or discontinued therapy due to lack of efficacy, adverse event (AE), or death were included in the safety population (n=642, Figure 1). Safety events are reported for up to 30 days following treatment completion. Death, liver transplantation, and development of HCC, were captured for all patients throughout the entire study period and were not restricted to 30 days post-treatment.
Adverse events (AEs) included any new, untoward events or exacerbated baseline conditions noted in the medical record. Abnormal lab values were captured as AEs if the abnormality was treated with a prescription therapy or required adjustment/discontinuation of HCV therapy. Serious adverse events (SAEs) were any AEs that required hospitalization or met criteria for expedited reporting per Food and Drug Administration form MEDWATCH 35000.
Hepatic decompensation during therapy was defined as experiencing one or more of the following: hepatic decompensation listed as an AE by a healthcare professional, new onset or exacerbation of hepatic encephalopathy, spontaneous bacterial peritonitis, variceal hemorrhage, ascites or hepatic hydrothorax, hepatorenal syndrome, hepatopulmonary syndrome as well as liver failure or if the patient received a new prescription for a medication to treat one of the aforementioned indications.
Impact of Treatment on Hepatic Function:
Changes in measures of hepatic function from baseline to either the short-term or the long-term follow-up time-point were calculated for all patients in the efficacy population who had sufficient follow-up time as well as availability of the laboratory data needed for MELD score calculation. Short-term follow-up was defined as the last available value recorded between 9-26 weeks after the end of treatment. Long-term follow-up was defined as the last available value recorded at least 36 weeks after the end of treatment. Hepatic function measures examined include total bilirubin, albumin, as well as the calculated MELD score. If a patient underwent liver transplant, all measures recorded at or after transplant were dismissed.
Statistical Analysis:
To assess treatment efficacy, unadjusted SVR12 rates with exact binomial confidence intervals were calculated for the overall efficacy population as well as by baseline MELD category (10–15, 16-20 and ≥21). To assess treatment effectiveness, SVR rates were also calculated for a subpopulation, which included the efficacy population and patients with no virologic outcomes because of death at or within one year after the end of treatment, with the latter considered ‘failures’. Associations between SVR12 and each of the most well-established baseline characteristics were estimated with univariable logistic regression. In an attempt to better understand the relationship among risk factors and odds of achieving SVR, we also performed multivariable analysis.
MELD score, total bilirubin and albumin were compared between baseline, the short-term follow-up time-point, and the long-term follow up time-point. Associations between short-term improvements in MELD of ≥3 points as well as achievement of final MELD <10, and baseline characteristics were also estimated. We selected at least a 3 MELD point improvement as clinically meaningful as it has been observed that intra-laboratory variability can have a significant impact on measured/calculated MELD, and most of this variability leads to a change in mean MELD of 3 points or less13. MELD analyses were performed excluding patients on dialysis at baseline to ensure that chronic kidney disease requiring dialysis did not have a significant impact on the trajectory of calculated MELD.
Time-to-event analysis was performed using the SAS PROC LIFETEST, for the events of death and liver transplant. Kaplan-Meier estimates of survivor functions were used for comparison of survival curves between groups of participants using a log-rank test. Analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC).
Results:
Patients:
A total of 667 patients with advanced/decompensated cirrhosis in HCV-TARGET were treated with DAA therapy in the study period. Patients who discontinued treatment early due to non-compliance (n=3), or administrative reasons (n=4), were lost to follow up/withdrew from study (n=15), underwent liver transplantation (n=2) or had records unavailable at the time of analysis (n=1) were not included in the cohort (Figure 1). The remaining 642 patients comprised the safety population. The overall median age was 60 years, 68% of patients were men and 68% Caucasian (Table 1). The majority of patients had HCV genotype 1 (70%), and 45% were treatment experienced.
The median MELD at the time of treatment initiation was 12 (range 10-39), with 521 (81%) with MELD 10-15, 78 (12%) MELD 16-20 and 43 (7%) MELD ≥ 21. Overall 309 patients (48.1%) had Child class B, 64% had experienced prior clinical decompensation events, and 4% of patients were on dialysis at treatment initiation.
The efficacy population (n=577) was restricted further by excluding patients who did not have a virologic outcome reported due to death (7 died during treatment and 26 died after treatment concluded but prior to evaluation of SVR, 21 of them within a year after treatment), and 32 who were lost to post-treatment follow-up.
Antiviral Treatment Regimen:
The most common DAA regimen was sofosbuvir/ledipasvir (56%), followed by sofosbuvir/velpatasvir (19%), sofosbuvir/daclatasvir (14%), elbasvir/grazoprevir (6%), glecaprevir/pibrentasvir (4%), and sofosbuvir/velpatasvir/voxilaprevir (2%). The most common regimen in patients with MELD ≥ 21 was elbasvir/grazoprevir, all of whom had a baseline GFR < 30 ml/min. Ribavirin was used in 31% (197/642) of patients in the safety population with 32% (169/521) in the MELD 10-15 group, 31% (24/78) in the MELD 16-20 group, and 1% (4/43) in the MELD ≥ 21 group. Median (interquartile range) treatment duration for the entire cohort was 90 days (85-169), and 91.0 days (85-169) among patients who completed treatment.
Treatment Response:
Of the 577 patients in the efficacy population, 522 (90.5%) achieved SVR12 and rates by MELD categories were: 427/475 (89.9%, 95%CI: 86.8-92.5) in MELD 10-15, 62/68 (91.2%, 95%CI: 81.8-96.7) in MELD 16-20, and 33/34 (97.1%, 95%CI: 84.7-99.9) in MELD ≥21 (Figure 2).
However, the efficacy population (defined as those with available virologic SVR status) does not include patients with no virologic outcome due to death (7 on and 21 within 1 year after the DAA treatment; 16 of these patients had MELD 10-15, 4 with MELD 16-20 and 8 with MELD ≥21.) To assess treatment effectiveness, we assigned virologic outcomes for these patients as ‘failures’ and calculated SVR rates by adding these 28 patients to the efficacy population defined above. SVR12 rates including these cases were statistically similar throughout all MELD categories: 427/491 (87.0%, 95%CI: 83.7-89.8) in MELD 10-15, 62/72 (86.1%, 95%CI: 75.9-93.1) in MELD 16-20 and 33/42 (78.6%, 95%CI: 63.2-89.7) in MELD ≥21 (Supplemental Figure 1).
In univariable analysis, negative predictors of SVR included the presence of ascites (OR 0.33, 95%CI 0.17-0.59, p<0.001), hepatocellular carcinoma (HCC)(OR 0.33, 95%CI 0.17-0.69, p=0.002), albumin level <3.5 mg/dl (OR 0.31, 95%CI 0.14-0.61, p=0.001) and proton pump inhibitor (PPI) use (OR 0.47, 95% CI 0.26-0.83, P=0.010), while treatment experience, baseline MELD and ribavirin use did not show association with treatment response. Positive predictors of SVR were total bilirubin <=1.2.mg/dl (OR 1.94, 95% CI 1.05-3.80, P=0.042) and Child-Pugh A (OR 4.33, 95% CI 1.90-12.04, P=0.002)(not shown).
The multivariable model with LASSO-selected covariates for association with SVR included the presence of ascites (OR 0.58, 95% CI 0.26-1.21, p=0.157), albumin level <3.5 mg/dl (OR 0.41, 95%CI 0.14-1.04, p=0.070), history of HCC (OR 0.40, 95% CI 0.18-0.91, p=0.024), PPI use (OR 0.59, 95% CI 0.31-1.08, p-value =0.099), age<60 (OR 0.58, 95% CI 0.30-1.09, p=0.085), male sex (OR 0.49, 95% CI 0.23-0.98, p=0.049), MELD<16 (OR 0.36, 95% CI 0.11-0.93, p=0.053), and Child-Pugh A (OR 1.84, 95% CI 0.53-7.28, p=0.342)(Supplemental Figure 2).
Short-Term Changes in Hepatic Function:
During short-term follow-up, MELD score decreased in 187 (56%), did not change in 54 (16%) and increased in 92 (28%) patients (Figure 3A). A decline of at least 3 MELD points occurred in 80 (24%) patients and was more common among patients with a baseline MELD ≥16 (41%)(Figure 3A). Short-term MELD change among patients who achieved SVR and who failed DAA treatment were similar (Figure 3B).
In multivariable model with LASSO-selected covariates, male sex (OR 0.51, 95% CI 0.27-0.93, p=0.029), baseline ALT<60 u/l (OR 0.41, 95% CI 0.22-0.75, p=0.004), Child-Pugh score A (OR 0.48, 95% CI 0.21-0.99, p=0.058) and MELD<16 (OR 0.33, 95% CI 0.16-0.69, p=0.003) were negatively associated with MELD improvement of at least 3 points, and age<60 (OR 2.15, 95% CI 1.20-3.93, p=0.012) was predictive of this improvement (Figure 4).
Long-Term Changes in Hepatic Function:
A total of 218 patients in the efficacy population who achieved SVR who were not on dialysis at baseline were evaluable for long-term outcomes. Of these, 158 had MELD scores available at least 36 weeks after the end of treatment. The median (interquartile range) follow-up after the end of treatment was 213 (174-224) weeks. Baseline demographics were similar between patients with and without long-term follow-up, including age (mean 58.5 vs. 59.8, p=0.168), HCC (8.2% vs 11.9%, p=0.290), baseline MELD (mean 13.2 vs. 12.9, p=0.417), and baseline albumin (mean 3.2 vs. 3.3, p=0.257) (Table 2). However, patients with long-term follow up were more likely to have ascites at baseline (55.7% vs. 44.1%, p=0.044) and prior hepatic decompensation events (71.5 % vs. 60.8 %, p=0.050).
The overall mean decrease in MELD score from the start of treatment until the end of long-term follow up was 0.30 (95% CI -1.10, 0.50) points. Patients with a baseline MELD ≥16 had a mean of 1.90 (95% CI -3.66, -0.15) points improvement in MELD, while those with a baseline MELD <16 had a mean increase of 0.09 (95% CI -0.80, 0.99)(Figure 5). A decline of at least 3 MELD points in long-term follow up occurred in 45 (29%) patients and was more common among those with a baseline MELD ≥16 (45 %). During long term follow up, 40 (25%) patients achieved a reduction in MELD score to <10, 38 of them with a baseline MELD 10-15 and 2 with a baseline MELD 16-20.
In univariable analysis, baseline MELD <16 (OR 5.08, 95% CI 1.57-25.81, p=0.020) and baseline INR <1.5 (OR 2.67, 95% CI 1.00-8.82 p=0.074) were predictive of achieving a final MELD score of <10 in long term follow up. The presence of ascites (OR 0.49, 95% CI 0.24-1.01, p=0,057) and ALT<60 u/l (OR 0.40, 95% CI 0.19-0.83, p=0.016) were negatively associated with this improvement, while age, gender, history of HCC and other baseline laboratory values did not show associations. We further evaluated CPT score and impact on long-term MELD outcome of <10 and noted that Child-Pugh A did not show an association with long-term MELD<10 (OR 1.55, 95% CI 0.64-3.59; p-value =0.324) while Child-Pugh C was negatively associated with long-term MELD<10 (OR 0.15, 95% CI 0.02-0.60; p-value =0.029). In multivariable analysis, LASSO regression identified baseline ALT<60 u/l (OR 0.39, 95% CI 0.17-0.86, p=0.022) and Child-Pugh C (OR 0.25, 95% CI 0.03-1.14, p=0.127) as negative predictors and MELD <16 (OR 2.75, 95% 0.76-14.7, p=0.165) as positive predictor for this improvement.
In long-term follow up, there was an overall increase in mean total bilirubin of 0.23 (95% CI -0.44, 0.89) mg/dl (Supplemental Figure 3). Patients with a baseline MELD of ≥16 had a mean decrease of 0.67 (95% CI -1.31, -0.03) mg/dl in total bilirubin, while those with a baseline MELD <16 experienced a mean increase in total bilirubin of 0.41 (95% CI -0.38, 1.20) mg/dl. In long-term follow-up there was also an increase in mean albumin of 0.36 (95% CI 0.28, 0.45) g/dl (Supplemental Figure 4). Patients with a baseline MELD of ≥16 had a mean improvement of 0.36 (0.14, 0.58) g/dl in albumin, and those with a baseline MELD <16 experienced a mean improvement of 0.36 (0.27, 0.46) g/dl.
Safety:
Overall, 468 (73%) patients of the safety population experienced an AE, leading to treatment discontinuation in 3.6%. SAEs occurred in 20% of the entire safety population, including 17% in MELD 10-15, 39% in MELD 16-20 and 28% in MELD ≥21 patients (Table 3).
The most common liver-related SAEs of interest were the development of hepatic encephalopathy (4.4%), ascites (0.9%), and bacterial peritonitis (0.8%). Overall, 83 patients (12.9%) had new decompensating events either on treatment or up to 30 days after the end of treatment and 41 patients (6.4%) had decompensating events that were classified as SAEs. Of the 83 patients, only 7 patients (8%) had not had any history of decompensating events reported prior to DAA regimen. 94 (69%) of the 137 AEs experienced by 61 (73%) of these 83 patients were exacerbations of previously recorded condition. 15 remaining patients (18%) with prior history of decompensating events had developed a different decompensating event than what was recorded prior to start of DAA treatment.
There was a total of 16 patients with baseline MELD ≥21 treated with elbasvir/grazoprevir. All of these patients had eGFR of <30 mg/dl and 15 of them were on dialysis at baseline. There were no decompensating events characterized as SAEs recorded in this subpopulation.
Survival:
Seven patients died on treatment (coronary artery disease, hepatic encephalopathy, hepatic failure, multi-organ failure, septic shock, and unknown cause in two patients), four of whom had a baseline MELD 10-15, one MELD 16-20 and two MELD ≥ 21. In addition, at least 44 patients died after the end of DAA treatment. Death was attributable to advanced liver disease in 12 patients, while the cause of death in the others was liver unrelated or unknown. Median duration (range) of time for survival follow-up was 15.5 (0–50.8) months. Baseline MELD <16 was predictive of overall survival (log rank, p<0.0001). One-, two-, and three- year survival rates in participants with baseline MELD 10-15 and baseline MELD ≥16 were 96.0%-93.6%-89.7% and 85.9%-82.4%-71.7%, respectively.
Five patients underwent liver transplantation while on treatment. Three achieved SVR, one failed treatment, and virologic data was unavailable in one patient. In addition, 17 patients underwent liver transplant after the end of treatment. Median (range) follow up time for liver transplant events was 14.1 (0–50.8) months. The proportion of patients with baseline MELD 10-15, and baseline MELD ≥16, who underwent liver transplant 1-, 2-, and 3 years after treatment initiation was 2.6%- 5.4%- 9.0% and 1.2%- 2.9%-2.9%, respectively. Median duration (range) of time for HCC diagnosis follow-up was 14.1 (0–50.8) months. Total of 13 patients with baseline MELD 10-15 and 3 patients with baseline MELD > 16 developed HCC.
| |
| |
| |
|
|
|
