Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma.

BACKGROUND/AIMS: Organ failure in patients with acute decompensation (AD) is a defining characteristic of acute-on-chronic liver failure (ACLF). However, the clinical features of AD during the long-term clinical course of hepatocellular carcinoma (HCC) are still poorly understood. This study aimed to clarify features and impact of AD/ACLF on the prognosis of patients after treatment for HCC.
METHODS: This retrospective study enrolled 556 consecutive patients who were initially diagnosed with HCC, and analyses were conducted taking into account HCC treatment type, HCC stage, and presence or absence of cirrhosis.
RESULTS: During follow-up, 299 patients with AD were hospitalized. AD occurrence is closely related to prognosis, regardless of the presence or absence of cirrhosis and HCC stage, and early-onset AD (within 90 days after HCC treatment) has negative impact on prognosis. In the intermediate-advanced-stage group, surgical resection had a positive impact on AD incidence post-treatment. After systemic therapy for HCC, renal impairment was the predictive factors for AD development. The 28/90-day mortality rate was higher among 41 cases (13.7%) with AD who exhibited ACLF as compared with cases without ACLF. AD without cirrhosis had similar ACLF incidence and short-term mortality, compared to AD with cirrhosis. The prognostic model using a decision-tree-based approach, which includes ACLF, bilirubin level, HCC progression, and MELD score is useful for predicting 90- or 28-day mortality after AD diagnosis.
CONCLUSIONS: Careful management of patients with HCC who are hospitalized with AD is necessary, considering ACLF, HCC progression, and liver function.

Introduction

Hepatocellular carcinoma (HCC) remains one of the leading causes of cancer-related death worldwide.1 Management of HCC should be based on proper assessment of disease severity, treatment, and surveillance [1-3]. Over the past few decades, despite the improvement of disease management, the outcomes of HCC remain unsatisfactory [1,4]. In patients with HCC, the prevalence of concomitant liver cirrhosis is greater than 80%, and the liver functional reserve is one of the critical factors affecting their prognosis [1-3].

The most common hospital presentation of patients with cirrhosis is acute decompensation (AD) with gastrointestinal bleeding, bacterial infection, acute onset of ascites, or hepatic encephalopathy (HE), alone or in combination [5-8]. The occurrence of hepatic and extrahepatic organ failure in patients with AD is indicative of acute-on-chronic liver failure (ACLF). ACLF occurs in approximately 30% of patients with acute AD and is associated with a 28-day mortality rate of roughly 30% [9-11].

However, the proposed definition of ACLF by the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Disease (AASLD) has been validated only in patients with cirrhosis [10]. In contrast, the Asian Pacific Association for the Study of the Liver (APASL) definition of ACLF includes chronic hepatitis regardless of the presence of cirrhosis [12], because chronic liver disease is difficult to clearly distinguish from cirrhosis [13,14]. In addition, most of the previous reports that demonstrated the impact of AD in patients with cirrhosis excluded the HCC population [10]; thus, the influence of AD on the clinical outcome after treatment for HCC and the influence of treatment for HCC on the incidence of AD or ACLF remains unclear.

We hypothesized that AD occurrence is a crucial determinant of the outcomes after treatment for HCC, regardless of the presence of cirrhosis. To validate this hypothesis and the clinical features of AD after treatment for HCC, we conducted the present study to elucidate the occurrence of AD during the long-term clinical course, and the influence of AD on the prognosis of patients after treatment for HCC. In addition, we evaluated the ACLF incidence and impact of AD on short-term prognosis after treatment for HCC, comparing HCC with cirrhosis versus without cirrhosis.

Patients and methods

Patients

This retrospective study included data obtained from our institutional database between October 2011 and December 2016. We enrolled consecutive patients who were initially diagnosed with HCC and scheduled for treatment. Patients receiving maintenance dialysis or who received only best supportive care were excluded. The treatment strategy for HCC was discussed at a multidisciplinary meeting. After the advantages and side effects of various therapies and recommendations from the experts were explained, patients finalized the treatment strategy.

This study conformed to the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Chiba University Graduate School of Medicine, and written informed consent was waived because of the retrospective design. Informed consent was obtained in the form of an opt-out on the web-site.

Definitions

HCC was diagnosed using the AASLD criteria (early-stage: single of any size or ≤ 3 nodules of ≤ 3 cm diameter; intermediate-stage: > 3 nodules of any size or 2–3 nodules of > 3 cm diameter; advanced-stage: any nodules with macrovascular invasion or extrahepatic spread). AD was defined according to the acute onset of ascites [15], HE [16], gastrointestinal/intra-abdominal hemorrhage [17], or bacterial infection [18], alone or in combination, requiring hospital treatment. ACLF was also defined with respect to organ failure according to the Chronic Liver Failure (CLIF)-Organ Failure (OF) score [9] and diagnosis required, in relation to known or unknown chronic liver disease. The CLIF-OF is scored as follows: grade 1 –(i) single kidney failure; (ii) single liver, coagulation, circulatory, or respiratory failure, and serum creatinine levels between 1.5 and 2 mg/dL and/or HE grade I or II; (iii) single cerebral failure (HE grade III or IV) associated with a serum creatinine between 1.5 and 2 mg/dL; grade 2, two organ failures; grade 3, three or more organ failures. Non-ACLF was defined as AD in patients who did not meet the criteria for ACLF diagnosis [9].

After we analyzed all non-ACLF patients at admission due to AD, the diagnosis of later-onset ACLF was reached. A patient was diagnosed with later-onset ACLF if the ACLF developed within 28 days after admission.

Evident cirrhosis was defined according to a combination of clinical signs and findings provided by laboratory tests, radiologic imaging, or liver biopsy. HE was assessed using the West Haven grading system [16]. The degree of ascites was defined according to international guidelines [15]: mild, ascites that was only detectable by ultrasound examination; moderate, ascites that caused moderate symmetrical distension of the abdomen; and severe, ascites that caused marked abdominal distension. Diagnosis of spontaneous bacterial peritonitis was confirmed when the ascitic neutrophil count was >250 cells/mm3 with no intra-abdominal and surgically treatable source of sepsis [15].

Statistical analysis

All data are expressed as the mean ± SD or as a percentage. Continuous variables were analyzed using Student’s t-test or Mann–Whitney U-test, as appropriate. Categorical variables were analyzed using Fisher’s exact test or chi-squared test, as appropriate. The multivariate analysis was assessed by logistic regression analysis. The cumulative survival rate was calculated using the Kaplan–Meier method. Risk factors for the development of AD were evaluated by Cox regression analysis. The best cut-off value was determined by receiver operating characteristics (ROC) analysis. Classification and regression tree (CART) analysis was performed using the R-powered data tool Exploratory (https::/exploratory.io) [19]. Statistical data were analyzed using SAS version 9.2 (SAS Institute, Cary, NC) and p < 0.05 was considered statistically significant.

Results

Patient characteristics

The study flowchart is depicted in Fig 1. Among 616 consecutive patients initially diagnosed with HCC during the study period, 556 were enrolled. Table 1 summarizes the clinical data for the study population. The median observation period was 36.7 months. Initially, 332 (59.7%) patients received potentially curative treatment (surgical resection or ablation). One patient received liver transplantation, and 223 patients died during the study period. The cumulative overall survival rates were 85.8% at 1 year and 66.8% at 3 years, and significantly lower in the intermediate-advanced–stage group (67.9% at 1 year, 36.1% at 3 years) than in the early-stage group (95.4% at 1 year, 83.8% at 3 years; p < 0.001).

Fig 1

Protocol diagram.

Table 1

Patient characteristics.

Number of patients	556
HCC early/intermediate-advanced stage	353/203
Age (years)	70 ± 9
Sex (male/female)	396/160
Liver cirrhosis	349 (62.7%)
Prior history of acute decompensation	97 (17.5%)
Etiology (virus/alcohol /NASH/PBC/AIH/cryptogenic)	339/71/34/8/4/100
Albumin-bilirubin (ALBI) score	-2.42 ± 0.49
Child-Pugh score	6 ± 1
Model for end-stage liver disease (MELD) score	7 ± 2
Alpha-fetoprotein >18 ng/ml*	272 (48.9%)
Treatment for HCC (surgical resection/ablation/chemoembolization/systemic therapy)	140/192/178/46

Data are presented as mean ± SD or number (%).

AIH, autoimmune hepatitis; HCC, hepatocellular carcinoma; NASH, nonalcoholic steatohepatitis; PBC, primary biliary cirrhosis.

*Median value of alpha-fetoprotein was 18 ng/ml.

Development of AD

During follow-up, 299 cases with AD were hospitalized. The overall AD incidence rates were 15.5% at 1 year and 28.2% at 3 years.

Table 2 shows predictive factors for the development of AD according to univariate and multivariate analyses. The best cut-off value for predicting AD development as determined by ROC analysis was 18 ng/ml for AFP. In the multivariate analysis, presence of evident liver cirrhosis (p < 0.001), intermediate-advanced stage HCC (p < 0.001), higher alanine aminotransferase level (p = 0.049), higher AFP level (p = 0,049), higher Child-Pugh score (p < 0.001), and cause of cirrhosis other than virus-related hepatitis (p = ≤ 0.001) were significant predictive factors for AD development.

Table 2

Cox regression analyses of predictive factors for the development of acute decompensation.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.991 (0.976–1.006)	0.235	-
Male sex	1.308 (0.945–1.811)	0.105	-
Liver cirrhosis	2.020 (1.459–2.796)	<0.001	1.843 (1.312–2.589)	<0.001
Prior history of acute decompensation	3.912 (2.821–5.426)	<0.001	-
Intermediate-advanced-stage HCC	3.749 (2.791–5.034)	<0.001	2.970 (2.154–4.095)	<0.001
Ascites	3.166 (2.186–4.584)	<0.001	-
Treatment for HCC
Surgical resection	0.459 (0.312–0.675)	<0.001	-
Ablation	0.544 (0.399–0.742)	<0.001	-
Chemoembolization	2.599 (1.942–3.478)	<0.001	-
Systemic therapy	4.192 (2.566–6.848)	<0.001	1.815 (1.071–3.075)	0.027
Etiology
Virus related hepatitis	0.532 (0.400–0.707)	<0.001	0.534 (0.396–0.719)	<0.001
Laboratory data
Alanine aminotransferase (U/L)	1.005 (1.002–1.008)	0.003	1.003 (1.000–1.006)	0.049
Bilirubin (mg/dL)	2.286 (1.801–2.902)	<0.001	-
Prothrombin time (international normalized ratio)	5.960 (2.904–12.232)	<0.001	-
Albumin (g/dL)	0.341 (0.254–0.456)	<0.001	-
Creatinine (mg/dL)	1.444 (0.909–2.294)	0.120	-
Sodium (mmol/L)	0.941 (0.903–0.982)	0.005	-
Platelets (109/L)	1.008 (0.989–1.028)	0.393	-
Alfa fetoprotein > 18 ng/ml*	1.869 (1.401–2.493)	<0.001	1.362 (1.001–1.853)	0.049
Albumin-bilirubin (ALBI) score	3.450 (2.542–4.682)	<0.001	-
Child-Pugh score	2.042 (1.788–2.333)	<0.001	1.746 (1.508–2.021)	<0.001
Model for end-stage liver disease (MELD) score	1.173 (1.102–1.248)	<0.001	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Predictive factors for the development of AD stratified with a type of treatment strategy for HCC

As shown in Tables 3–6, Intermediate-advanced stage HCC was significantly associated with the development of AD after all treatment strategies other than systemic therapy. Cause of cirrhosis other than virus-related hepatitis was the significant predictive factor for the development of AD after potentially curative treatment (surgical resection group, p = 0.043; ablation group, p = 0.006). Prior history of AD (ablation group, p = 0.018), higher AFP level (ablation group, p = 0.011), higher ALBI score (ablation group, p = 0.003), higher Child-Pugh score (chemoembolization group, p < 0.001; systemic therapy group, p < 0.001), higher alanine aminotransferase level (chemoembolization group, p = 0.009), higher bilirubin level (chemoembolization group, p = 0.014), and, higher creatinine level (systemic therapy group, p = 0.007) were the predictive factors for the development of AD after each type of treatment strategy for HCC.

Table 3

Cox regression analyses of predictive factors for the development of acute decompensation in the surgical resection group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	1.022 (0.983–1.063)	0.266	-
Male sex	1.111 (0.496–2.491)	0.798	-
Liver cirrhosis	1.769 (0.871–3.591)	0.114	-
Prior history of acute decompensation	2.508 (0.337–18.634)	0.369	-
Ascites	0.000 (0.000-)	0.990	-
Intermediate-advanced-stage HCC	2.683 (1.235–5.828)	0.013	2.733 (1.256–5.947)	0.011
Etiology
Virus related hepatitis	0.475 (0.227–0.992)	0.048	0.467 (0.223–0.997)	0.043
Laboratory data
Alanine aminotransferase (U/L)	1.009 (0.999–1.019)	0.069	-
Bilirubin (mg/dL)	0.620 (0.210–1.824)	0.385	-
Prothrombin time (international normalized ratio)	1.963 (0.233–16.520)	0.535	-
Albumin (g/dL)	0.956 (0.425–2.153)	0.914	-
Creatinine (mg/dL)	2.650 (0.625–11.237)	0.186	-
Sodium (mmol/L)	0.978 (0.841–1.136)	0.767	-
Platelets (109/L)	1.011 (0.953–1.073)	0.713	-
Alfa fetoprotein > 18 ng/ml*	1.850 (0.910–3.760)	0.089	-
Albumin-bilirubin (ALBI) score	0.970 (0.386–2.436)	0.949	-
Child-Pugh score	0.921 (0.387–2.192)	0.852	-
Model for end-stage liver disease (MELD) score	1.142 (0.906–1.438)	0.260	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Table 6

Cox regression analyses of predictive factors for the development of acute decompensation in the systemic therapy group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.987 (0.950–1.026)	0.509	-
Male sex	- (0.000-)	0.993	-
Liver cirrhosis	2.181 (0.815–5.842)	0.121	-
Prior history of acute decompensation	2.551 (0.963–6.757)	0.060	-
Ascites	2.020 (0.780–5.234)	0.148	-
Intermediate-advanced-stage HCC	1.840 (0.242–13.996)	0.556	-
Etiology
Virus related hepatitis	0.497 (0.194–1.277)	0.146	-
Laboratory data
Alanine aminotransferase (U/L)	1.001 (0.988–1.014)	0.886	-
Bilirubin (mg/dL)	1.827 (1.100–3.035)	0.020	-
Prothrombin time (international normalized ratio)	3.866 (1.054–14.184)	0.042	-
Albumin (g/dL)	0.189 (0.065–0.549)	0.002	-
Creatinine (mg/dL)	20.510 (2.509–167.658)	0.005	12.211 (1.973–75.576)	0.007
Sodium (mmol/L)	1.035 (0.858–1.249)	0.716	-
Platelets (109/L)	1.028 (0.986–1.072)	0.194	-
Alfa fetoprotein > 18 ng/ml*	1.378 (0.315–6.036)	0.670	-
Albumin-bilirubin (ALBI) score	6.086 (2.079–17.816)	0.001	-
Child-Pugh score	2.297 (1.475–3.579)	<0.001	2.412 (1.488–3.909)	<0.001
Model for end-stage liver disease (MELD) score	1.309 (1.114–1.540)	0.001	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Predictive factors for the development of AD stratified with HCC stage

Tables 7 and 8 show the results of predictive factors for the development of AD stratifying HCC into early-stage HCC and intermediate-advanced-stage HCC. In early-stage HCC group, presence of evident liver cirrhosis (p < = 0.001), prior history of AD (p < 0.001), higher Child-Pugh score (p < 0.001), and cause of cirrhosis other than virus-related hepatitis (p = 0.001) were significant predictive factors for AD development. In intermediate-advanced-stage HCC group, advanced-stage HCC (p = 0.011), higher bilirubin level (p = 0.033), and higher Child-Pugh score (p < 0.001) were significant predictive factors for AD development, and surgical resection for HCC was negatively associated with AD development (p = 0.007).

Table 7

Cox regression analyses of predictive factors for the development of acute decompensation in the early-stage group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.998 (0.976–1.020)	0.853	-
Male sex	0.845 (0.562–1.272)	0.420	-
Liver cirrhosis	2.902 (1.756–4.797)	<0.001	2.373 (1.402–4.016)	0.001
Prior history of acute decompensation	3.945 (2.379–6.542)	<0.001	1.604 (1.260–2.043)	<0.001
Ascites	2.233 (1.158–4.307)	0.017	-
Treatment for HCC
Surgical resection	0.577 (0.356–0.937)	0.026	-
Ablation	1.019 (0.682–1.524)	0.926	-
Chemoembolization	2.058 (1.254–3.380)	0.004	-
Systemic therapy	6.432 (0.876–47.251)	0.067	-
Etiology
Virus related hepatitis	0.522 (0.350–0.779)	0.002	0.504 (0.336–0.755)	0.001
Laboratory data
Alanine aminotransferase (U/L)	1.002 (0.996–1.008)	0.468	-
Bilirubin (mg/dL)	2.240 (1.484–3.380)	<0.001	-
Prothrombin time (international normalized ratio)	4.511 (1.078–18.872)	0.039	-
Albumin (g/dL)	0.350 (0.260–0.607)	<0.001	-
Creatinine (mg/dL)	1.532 (0.701–3.351)	0.285	-
Sodium (mmol/L)	0.958 (0.901–1.019)	0.175	-
Platelets (109/L)	0.942 (0.908–0.978)	0.002	-
Alfa fetoprotein > 18 ng/ml*	1.515 (1.015–2.261)	0.042	-
Albumin-bilirubin (ALBI) score	3.379 (2.166–5.272)	<0.001	-
Child-Pugh score	1.936 (1.550–2.419)	<0.001	1.604 (1.260–2.043)	<0.001
Model for end-stage liver disease (MELD) score	1.181 (1.054–1322)	0.004	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Table 8

Cox regression analyses of predictive factors for the development of acute decompensation in the intermediate-advanced-stage group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.993 (0.973–1.014)	0.509	-
Male sex	1.508 (0.804–2.829)	0.201	-
Liver cirrhosis	1.548 (1.000–2.397)	0.050	-
Prior history of acute decompensation	2.377 (1.535–3.682)	<0.001	-
Ascites	2.617 (1.631–4.200)	<0.001	-
HCC advanced stage	1.630 (1.080–2.459)	0.020	1.780 (1.140–2.780)	0.011
Treatment for HCC
Surgical resection	0.413 (0.214–0.797)	0.008	0.378 (0.186–0.768)	0.007
Ablation	0.883 (0.279–2.796)	0.833	-
Chemoembolization	1.198 (0.779–1.842)	0.411	-
Systemic therapy	1.950 (1.151–3.305)	0.013	-
Etiology
Virus related hepatitis	0.660 (0.439–0.992)	0.046	-
Laboratory data
Alanine aminotransferase (U/L)	1.003 (0.999–1.066)	0.125	-
Bilirubin (mg/dL)	2.191 (1.651–2.908)	<0.001	1.484 (1.032–2.134)	0.033
Prothrombin time (international normalized ratio)	4.024 (1.790–9.046)	0.001	-
Albumin (g/dL)	0.406 (0.263–0.626)	<0.001	-
Creatinine (mg/dL)	1.041 (0.591–1.835)	0.889	-
Sodium (mmol/L)	0.955 (0.899–1.014)	0.135	-
Platelets (109/L)	1.013 (0.993–1.033)	0.208	-
Alfa fetoprotein > 18 ng/ml*	1.393 (0.896–2.165)	0.141	-
Albumin-bilirubin (ALBI) score	2.922 (1.889–4.521)	<0.001	-
Child-Pugh score	1.957 (1.631–2.347)	<0.001	1.592 (1.261–2.009)	<0.001
Model for end-stage liver disease (MELD) score	1.111 (1.029–1.200)	0.007	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Impact of AD on prognosis

Cumulative overall survival was significantly lower in patients with the occurrence of AD than those without (without AD vs. with AD: 93.2% vs. 76.9% at 1 year/78.9% vs. 48.5% at 3 years, p < 0.001). When stratified AD into early-onset AD (within 90 days after treatment for HCC) and later-onset AD (after 90 days of treatment for HCC), early-onset AD negatively affected long-term survival, compared to later-onset AD (non-AD vs. later-onset AD vs. early-onset AD: 90.8% vs. 86.7% vs. 33.8% at 1 year/78.9% vs. 55.7% vs. 16.9% at 3 years, p < 0.001).

Table 9 shows the prognostic factors. Multivariate analysis identified the following significant prognostic factors: the prior history of AD (p = 0.001), intermediate-advanced HCC (p < 0.001), receiving potentially curative treatment as the primary treatment (p = 0.001), the occurrence of AD during clinical course (p < 0.001), higher AFP level (p = 0.004), and higher Child-Pugh score (p < 0.001).

Table 9

Cox regression analyses of prognostic factors.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.984 (0.971–0.998)	0.021	-
Male sex	1.623 (1.174–2.244)	0.003	-
Liver cirrhosis	1.597 (1.190–2.143)	0.002	-
Prior history of acute decompensation	4.270 (3.175–5.744)	<0.001	1.839 (1.294–2.613)	0.001
Ascites	4.611 (3.555–6.336)	<0.001	-
Intermediate-advanced-stage HCC	4.851 (3.672–6.408)	<0.001	2.547 (1.792–3.620)	<0.001
Acute decompensation during clinical course	3.985 (2.950–5.141)	<0.001	2.371 (1.769–3.177)	<0.001
Treatment for HCC
Potentially curative treatment	0.217 (0.164–0.287)	<0.001	0.546 (0.380–0.785)	0.001
Etiology
Virus related hepatitis	0.635 (0.485–0.831)	0.001	-
Laboratory data
Alanine aminotransferase (U/L)	1.005 (1.002–1.008)	<0.001	-
Bilirubin (mg/dL)	1.747 (1.378–2.216)	<0.001	-
Prothrombin time (international normalized ratio)	4.996 (2.425–10.295)	<0.001	-
Albumin (g/dL)	0.373 (0.284–0.490)	<0.001	-
Creatinine (mg/dL)	1.390 (0.883–2187)	0.155	-
Sodium (mmol/L)	0.944 (0.908–0.983)	0.005	-
Platelets (109/L)	1.022 (1.006–1.039)	0.008	-
Alfa fetoprotein > 18 ng/ml*	2.228 (1.691–2.936)	<0.001	1.524 (1.147–2.024)	0.004
Albumin-bilirubin (ALBI) score	2.907 (2.193–3.855)	<0.001	-
Child-Pugh score	1.997 (1.763–2.262)	<0.001	1.449 (1.242–1.690)	<0.001
Model for end-stage liver disease (MELD) score	1.130 (1.058–1.207)	<0.001	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

AD occurrence was associated with the poor prognosis after treatment for HCC, regardless of the presence or absence of cirrhosis (without AD vs with AD: absence of cirrhosis, 93.9% vs. 72.0% at 1 year/82.3% vs. 51.6% at 3years, p < 0.001; presence of cirrhosis, 88.2% vs. 78.5% at 1 year/77.6% vs. 47.3% at 3 years, p < 0.001) and the stage of HCC (without AD vs with AD: early-stage, 96.4% vs. 92.7% at 1 year/90.6% vs. 67.3% at 3years, p < 0.001; intermediate-advanced-stage, 76.0% vs. 59.7% at 1 year/45.3% vs. 26.7% at 3 years, p < 0.001).

Development of ACLF

Among 299 cases with AD, 41 (13.7%) had ACLF at admission due to AD. In cases with and without ACLF (non-ACLF), the mortality rates were 68.3% versus 19.4% at 28 days after admission (p < 0.001) and 87.8% versus 41.5% at 90 days after admission (p < 0.001). The ACLF grade was associated with the short-term prognosis (28/90-day mortality: non-ACLF, 19.4%/41.5%; grade 1, 61.5%/92.0%; grades 2–3, 80.0%/86.7%, p < 0.001, respectively). Of the 258 non-ACLF cases at admission, 23 (8.9%) developed ACLF within 28 days (later-onset ACLF), and their 28- and 90-day mortality rates were significantly higher than patients who did not develop later-onset ACLF (15.7% vs. 56.5% at 28 days, p < 0.001; 37.0% vs. 87.0% at 90 days, p < 0.001).

Comparing AD with cirrhosis and those without, there is no significant difference in the ACLF incidence (with cirrhosis vs. without cirrhosis: 14.5% vs. 13.5%, p = 0.830), later-onset ACLF incidence (10.6% vs. 3.4%, p = 0.090), and 28-/90-day mortality (27.4% vs. 21.7% at 28 days, p = 0.348; 47.8% vs. 47.8% at 90 days, p = 1.000).

Prognostic model for 28/90-day mortality in cases with AD

Table 10 lists the predictive factors for 28/90-day mortality, according to univariate analysis, among 299 cases with AD. In multivariate analysis, the significant predictive factors for 28-day mortality were AD with ALCF (odds ratio [OR], 6.510; 95% confidence interval [CI], 2.844–14.900; p < 0.001), HCC progression as a potential precipitating event (OR, 3.842; 95% CI, 1.846–7.998; p < 0.001), white blood cell level (OR, 1.079; 95% CI, 1.023–1.137; p = 0.005), bilirubin level (OR, 1.140; 95% CI, 1.056–1.231; p = 1.140), albumin level (OR, 0452; 95% CI, 0.237–0.862; p = 0.016), and bacterial infection as a complication defining AD (OR, 0.305; 95% CI, 0.144–0.647; p = 0.002). The significant predictive factors for 90-day mortality were presence of ascites (OR, 3.133; 95% CI, 1.531–6.410; p = 0.002), HCC progression as a potential precipitating event (OR, 5.561; 95% CI, 2.566–12.052; p < 0.001), chemoembolization as a potential precipitating event (OR, 0.228, 95% CI, 0.055–0.941, p = 0.041), white blood cell level (OR, 1.087; 95% CI, 1.026–1.152; p = 0.005), albumin level (OR, 0534; 95% CI, 0.296–0.965; p = 0.038), MELD score (OR, 1.143; 95% CI, 1.079–1.211; p < 0.001), and ascites as a complication defining AD (OR, 0.296; 95% CI 0.135–0.648; p = 0.002). To construct a prognostic model, the most relevant variables were chosen for CART analysis, and several trees were constructed using an exploratory strategy. The final selected tree-discriminated cases were classified according to the following 3 subpopulations with distinct prognoses: prognostic model for 28-day mortality, low risk (Non-ACLF and bilirubin level < 9 mg/dL; 28-day mortality, 16%), intermediate risk (Non-ACLF and bilirubin ≥ 9 mg/dL; 28-day mortality, 61%), and high risk (AD with ACLF; 28-day mortality, 68%); and prognostic model for 90-day mortality, low risk (MELD score < 18 and without HCC progression as a potential precipitating event; 90-day mortality, 30%), intermediate risk (MELD score < 18 and with HCC progression as a potential precipitating event; 90-day mortality, 68%) and high risk (MELD score ≥ 18; 90-day mortality, 86%).

Table 10

Patient characteristics stratified by survivors or non-survivors at day 28 and day 90 in cases with acute decompensation.

	28-day			90-day
	Survivors (N = 221)	Non-survivors (N = 78)	P value	Survivors (N = 156)	Non-survivors (N = 143)	P value
HCC intermediate-advanced stage	91 (41.2%)	43 (55.1%)	0.003	61 (39.1%)	73 (51.1%)	0.038
ACLF	13 (5.9%)	28 (35.9%)	<0.001	5 (3.2%)	36 (25.2%)	<0.001
Age (years)	72 ± 8	68 ± 10	0.003	72 ± 8	70 ± 9	0.008
Male	155 (70.1%)	61 (78.2%)	0.171	105 (67.3%)	111 (77.6%)	0.047
Liver cirrhosis	167 (75.6%)	63 (80.8%)	0.348	120 (76.9%)	110 (76.9%)	1.000
Prior history of acute decompensation	62 (28.1%)	28 (35.9%)	0.194	45 (28.9%)	45 (31.5%)	0.621
Ascites	138 (62.4%)	66 (84.6%)	<0.001	87 (55.8%)	117 (81.8%)	<0.001
Etiology
Virus related hepatits	100 (45.3%)	35 (44.9%)	0.954	74 (47.4%)	61 (42.7%)	0.407
Complications defining AD
Ascites	48 (21.7%)	13 (16.7%)	0.341	39 (25.0%)	22 (15.4%)	0.039
Hepatic encephalopathy	26 (11.8%)	18 (23.1%)	0.015	18 (11.5%)	26 (18.2%)	0.105
Hemorrhage	57 (25.8%)	28 (35.9%)	0.089	41 (26.3%)	44 (30.8%)	0.390
Bacterial infection	90 (40.7%)	19 (24.4%)	0.010	58 (37.2%)	51 (35.7%)	0.786
HCC related potential precipitating events
HCC progression	29 (13.1%)	29 (37.2%)	<0.001	13 (8.3%)	45 (31.5%)	<0.001
Systemic chemotherapy	6 (2.7%)	3 (3.9%)	0.615	5 (3.2%)	4 (2.8%)	0.837
Chemoembolization	17 (7.7%)	2 (2.6%)	0.110	15 (9.6%)	4 (2.8%)	0.016
Surgical resection/ablation	5 (2.3%)	2 (2.6%)	0.880	3 (1.9%)	4 (2.8%)	0.618
Laboratory data
Alanine aminotransferase (U/L)	61 ± 122	97 ± 130	0.034	66 ± 142	76 ± 104	0.457
White blood cell (x109/L)	7.9 ± 5.0	10.6 ± 7.3	0.004	7.4 ± 5.0	10.0 ± 6.3	<0.001
Bilirubin (mg/dL)	2.6 ± 2.9	5.6 ± 6.0	<0.001	2.3 ± 2.5	4.6 ± 5.1	<0.001
Prothrombin time (international normalized ratio)	1.24 ± 0.30	1.39 ± 0.34	<0.001	1.22 ± 0.30	1.35 ± 0.32	0.001
Albumin (g/dL)	2.8 ± 0.5	2.5 ± 0.5	<0.001	2.9 ± 0.5	2.6 ± 0.5	<0.001
Creatinine (mg/dL)	1.04 ± 0.53	1.52 ± 0.84	<0.001	0.96 ± 0.45	1.39 ± 0.77	<0.001
Sodium (mmol/L)	135 ± 5	134 ± 7	0.294	136 ± 4	133 ± 6	<0.001
Platelets (109/L)	137 ± 99	163 ± 120	0.092	131 ± 99	158 ± 111	0.026
CLIF-C AD score	53.8 ± 8.7	60.5 ± 8.2	<0.001	52.0 ± 8.6	59.4 ± 8.0	<0.001
ALBI score	-1.39 ± 0.54	-0.94 ± 0.55	<0.001	-1.47 ± 0.51	-1.05 ± 0.57	<0.001
Child-Pugh score	8 ± 2	10 ± 2	< 0.001	8 ± 2	9 ± 2	<0.001
MELD score	11 ± 5	18 ± 7	<0.001	10 ± 4	16 ± 7	<0.001

Data are expressed as mean ± SD or number (%).

ACLF, acute-on-chronic liver failure; AD, acute decompensation; ALBI, albumin-bilirubin; CLIF-C AD, chronic liver failure-consortium acute decompensation; HCC, hepatocellular carcinoma; MELD, model for end-stage liver disease.

Discussion

This study determined the incidence of AD/ACLF and the risk factors for the occurrence of AD during the long-term clinical course after treatment for HCC, stratified with a type of HCC strategy and HCC stage. Our results suggested that AD occurrence was closely related with the prognosis after treatment for HCC, regardless of the presence or absence of cirrhosis and the stage of HCC, and early-onset AD had another impact on prognosis. Furthermore, this study showed that AD without cirrhosis had similar ACLF incidence and similar short-term mortality after admission due to AD, compared to AD with cirrhosis. The prognostic model using a decision-tree–based approach, including ACLF, bilirubin level, HCC progression, and MELD score, was useful for predicting 90/28-day mortality after AD diagnosis.

The AD incidence (15.5%) at 1 year was higher in this study than that reported in the literature [20,21]. This finding is consistent with previous studies in which the incidence of decompensation was higher in patients with HCC than in those without HCC [22,23]. Decreased liver function is crucial for AD development. Some treatments for HCC [24,25], and the development of HCC itself, have a negative impact on liver function. Recently, surgery has been described as a trigger for ACLF in patients with AD because of the invasiveness of the procedure [26]. However, we showed that surgical resection had a positive impact on AD incidence in the intermediate-advanced–stage group. The underlying reason might be that surgery is likely to be selected for patients with better liver function and less severe HCC. In fact, patients who received surgery had lower Child-Pugh scores than those who underwent non-surgical treatment in this study (p < 0.001), and had no prior history of AD. This suggests that for patients with preserved liver function, surgery might be a treatment option even in cases with intermediate-advanced–stage HCC group. In addition, virus-infected patients showed a lower rate of AD incidence, especially in the group which received ablation or surgical resection, possibly because of improvements in HCV/HBV treatment. In fact, 57% of patients with virus-related HCC were treated with antivirals, and there was a significant difference in the development of AD between treated and non-treated patients (treated vs. non-treated: 5.4% vs. 17.6% at 1 year/12.7% vs. 42.4% at 3 years, p < 0.001). Apart from liver function, during systemic therapy for HCC, patients with renal impairment should be monitored carefully because of high risk of AD development. In terms of long-term mortality, cirrhosis was not a significant prognostic factor in our study. However, prior history of AD and AD occurrence were associated with prognosis. In addition, early-onset AD after HCC treatment had negative impact on prognosis, compared to later-onset AD. These results indicate that AD, rather than cirrhosis, represents poor liver function and has negative impact on prognosis.

Notably, ACLF incidence in our study was relatively low, reaching 14% [8,27], but this may be because patients with HCC undergo stricter follow-up and are managed more carefully. Another possible reason is that ACLF incidence was around 21% when later-onset ACLF was included in the analysis, comparable to previous reports [10,27,28]. Moreover, even when cases of later-onset ACLF were excluded, the short-term mortality rate of AD without ACLF was very high compared with the reported short-term mortality rates for AD and ACLF [10,13]. One of the probable causes of the high mortality rate of AD is cancer cachexia caused by HCC, considering the result of high short-term mortality in AD caused by HCC progression. Another possible explanation is that cancer-related inflammation may be strongly related to the high short-term mortality rate of AD in patients with HCC, which might be supported by the result of poor short-term prognosis in AD with elevated white blood cell level. Inflammation caused by viral hepatitis, alcoholic hepatitis, and NASH leads to hepatocyte death and the repetitive division and proliferation of hepatocytes (so-called compensatory proliferation), which results in genetic mutations caused by DNA replication errors and the development of fibrosis and HCC [29]. Aside from the background inflammatory state of the liver, malignant transformation of cells induces an aberrant functional response to produce proinflammatory mediators in the tumor microenvironment, even without prior causative inflammation, and triggers the expression of inflammatory mediators such as cytokines and chemokines that amplify the inflammatory symptoms [29-31]. In contrast, bacterial infection as a complication defining AD had positive impact on short-term prognosis. This may be because patients with HCC tend to be treated promptly with antibiotics, which lead to early control of bacterial infection.

As to whether or not to include chronic hepatitis cases in the ACLF diagnosis, our study showed that cirrhosis was associated with AD development, but the short-term mortality was not significantly different between chronic hepatitis and cirrhosis cases. The results were also similar in cases with a history of decompensation, consistent with previous reports in which cases without a history of AD had higher levels of inflammatory mediators than those with a history of AD [10]. Further study is needed to determine whether chronic hepatitis cases should be included in the ACLF diagnosis. However, even in chronic hepatitis cases, careful management of AD in patients after treatment for HCC is necessary.

Similar to the conclusion of previous studies [32,33], all of the scoring systems used in the present study, including the Child–Pugh, ALBI, CLIF-C AD, and MELD scores, correlated well with the prognosis in the present study. In addition, given the high short-term mortality rate, liver transplantation might be considered in cases with HCC who meet MELD score ≥ 18 and Milan criteria.

Our study had several limitations. First, the data were retrospectively analyzed. Second, we excluded patients undergoing dialysis because of the problem of AD events in dialysis cases being diagnosed as ACLF. Therefore, the concept of ACLF in patients undergoing dialysis needs to be re-examined in a large cohort.

In conclusion, the occurrence of AD in patients with HCC correlates with a high mortality rate regardless of the stage of cirrhosis or liver cancer. In addition, patients with HCC who are hospitalized with AD should be carefully managed even if they have not reached the ACLF stage. Moreover, patients with impaired renal function who receive systemic therapy should be monitored carefully because of high incidence of AD. Lastly, a basic model using a simple CART algorithm is useful for estimating the prognosis for short-term mortality.

6 Aug 2021

PONE-D-21-23680

Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma

PLOS ONE

Dear Dr. Kondo,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

While both reviewers appreciated the usefulness of your dataset, they both have difficulties with the structure of the manuscript, a clear definition of hypothesis etc. This issues need to be improved before the manuscript can be considered further.

Please submit your revised manuscript by Sep 20 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Pavel Strnad

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I have read the article of Kondo et al “Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma” with great interest.

It describes an important topic on prognosis of HCC patients with HCC and the impact of acute decompensation. The aim of the study was to evaluate features and the impact of AD on the prognosis of HCC patients. The authors conducted a large retrospective study with 563 patients diagnosed with HCC, stratified them into early- and intermediate/advance group and measured occurrence of AD and ACLF episodes. The authors use internationally approved definitions and statistics are adequate.

The major comment I have is that the hypothesis/aim is somewhat confusing. On one hand they state that most studies that evaluated impact of AD in cirrhosis excluded HCC patients, and therefore the impact of AD on HCC patients is unknown. On the other hand, the authors hypothesize that AD occurrence is a crucial determinant of outcomes in HCC patients, regardless of cirrhosis. This are two important different questions. The first one requires a comparison between HCC cirrhotic patients and cirrhotic patients without HCC, and compare the occurrence of AD and the impact on survival. The second question requires a comparison between HCC patients with and without cirrhosis, and evaluate the occurrence on AD and prognosis.

By performing the analyses on the whole group (including HCC with cirrhosis and without cirrhosis) but without the comparison to a group without HCC, it is difficult to interpret these results. The main finding is that AD/ACLF occurrence is related to prognosis, but this is also the case in cirrhotic patients without HCC.

In addition, different treatment strategies were used, with potential different impacts on possible occurrence of AD. How did the authors correct for this effect, it seems important to me to separate the early complications from the potential beneficial effect in the long term when evaluating the exact impact of HCC and treatment on AD occurrence? I can imagine that an early complication presenting as AD has another impact on prognosis than an AD episode occurring years after the treatment. Moreover, the characteristics of patients that are treated with chemoembolization are completely different that patients with surgical resection or systemic therapy, it is difficult to compare these patient groups directly. I can imagine that an episode of AD impact survival differently in systemic therapy (with a HCC still present in the liver) and after resection (without an HCC). Ideally, these groups should be analyzed separately.

Finally, I’m missing AD occurrence (as a total group) in the prognostic models, given the aim of the study this would also be interesting to evaluate. Why did the authors stratify for early and advanced stage in prognostic modelling? It is also possible to include this variable as a predictive in the model, this would also results in larger groups which increases the power of the analysis.

In conclusion, while the authors have a very interesting dataset, the current analyses do not adequately answer the aims of the study. My suggestion is to remove most of the underpowered subgroup analyses and to focus on a main question that they want to answer, for example occurrence of AD after HCC treatment and impact on prognosis (but then results should be presented separately per treatment and take into account the different HCC populations undergoing these treatments). In the current form it is difficult to place these results in clinical context.

Minor comments:

- Why did the authors exclude Child-pugh C patients? Or patients with performance status of 2-3?

- Was liver biopsy known in every patient? The methods sections suggests this but it is not standard to biopsy every cirrhotic patient

- The numbers of the ACLF group are too low to perform prognostic modelling.

Reviewer #2: Kondo et al. performed a retrospective study to evaluate the impact of AD and ACLF on outcome in patients with HCC. Authors were trying to identify independent factors impacting on prognosis in this cohort.

Although this topic is interesting it is a difficult to read paper mostly due to the sequence results are presented. Different types of analysis (frequency of events, prognosis, univariate analysis, multivariate analysis) and distinct subgroups are put together into paragraphs which makes it difficult to comprehend the final messages.

General aspects:

I suggest, and this is my understanding of the manuscript, that authors make sure to address the following aspects and to put them in a more logical order:

• What is the relevance of AD and ACLF in cirrhosis with HCC (frequency and impact on prognosis)

• Is HCC and/or the stage of liver disease a factor impacting on the frequency of AD/ACLF (e.g. frequency and impact of HCC treatment related AD/ACLF, comparison with non-treatment related AD/ACLF; how many HCC related and how many AD/ACLF related deaths etc).

• Are there further independent factors impacting on patients’ outcome?

More specific aspects:

I am also missing the justification for dividing the groups upfront into early and late stage HCC. Although it is logical from the oncological point of view this paper is trying to elaborate on the decompensated cirrhosis/ACLF, which is meant to be biologically different. Authors should explain why they chose to do this subgroup analysis. I understand the results as if there are many data suggesting the severity of liver disease rather than the HCC itself as the main determinant for risk of AD/ACLF and subsequent outcome.

Why did authors include patients without cirrhosis? EASL CLIF ACLF definition requires the presence of liver cirrhosis.

At baseline, did all patients included have compensated cirrhosis?

Please explain the rational for excluding Child-C patients?

How many patients were transplanted during follow up? I understood that eligibility for transplant was no exclusion criterion.

Authors use the terms admission and hospitalisation – is that meant to be a surrogate for admission due to AD/ACLF? It is a bit confusing as admissions in general are not always due to AD/ACLF.

What is meant by curative treatment – intended to be or ultimately curative?

How many patients developed AD/ACLF as the consequence of HCC therapy? Are there factors impacting on AD/ACLF after HCC therapy?

Were all patients with viral hepatitis treated with antivirals? Any information on viral load etc? Were there differences between treated and non-treated patients?

There are too many variables in the regression analysis. For treatment and aetiology it may be useful to have only a binary variable (categorical), e.g. treatment/no treatment or viral hepatitis/others

Table 5 – Ascites, HE etc…. are no precipitating events, but complications of cirrhosis defining decompensation.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

23 Oct 2021

PONE-D-21-23680

Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma

PLOS ONE

To the Reviewers

Thank you for your suggestions and critical review.

Reviewer #1: I have read the article of Kondo et al “Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma” with great interest.

It describes an important topic on prognosis of HCC patients with HCC and the impact of acute decompensation. The aim of the study was to evaluate features and the impact of AD on the prognosis of HCC patients. The authors conducted a large retrospective study with 563 patients diagnosed with HCC, stratified them into early- and intermediate/advance group and measured occurrence of AD and ACLF episodes. The authors use internationally approved definitions and statistics are adequate.

The major comment I have is that the hypothesis/aim is somewhat confusing. On one hand they state that most studies that evaluated impact of AD in cirrhosis excluded HCC patients, and therefore the impact of AD on HCC patients is unknown. On the other hand, the authors hypothesize that AD occurrence is a crucial determinant of outcomes in HCC patients, regardless of cirrhosis. This are two important different questions. The first one requires a comparison between HCC cirrhotic patients and cirrhotic patients without HCC, and compare the occurrence of AD and the impact on survival. The second question requires a comparison between HCC patients with and without cirrhosis, and evaluate the occurrence on AD and prognosis.

By performing the analyses on the whole group (including HCC with cirrhosis and without cirrhosis) but without the comparison to a group without HCC, it is difficult to interpret these results. The main finding is that AD/ACLF occurrence is related to prognosis, but this is also the case in cirrhotic patients without HCC.

In addition, different treatment strategies were used, with potential different impacts on possible occurrence of AD. How did the authors correct for this effect, it seems important to me to separate the early complications from the potential beneficial effect in the long term when evaluating the exact impact of HCC and treatment on AD occurrence? I can imagine that an early complication presenting as AD has another impact on prognosis than an AD episode occurring years after the treatment. Moreover, the characteristics of patients that are treated with chemoembolization are completely different that patients with surgical resection or systemic therapy, it is difficult to compare these patient groups directly. I can imagine that an episode of AD impact survival differently in systemic therapy (with a HCC still present in the liver) and after resection (without an HCC). Ideally, these groups should be analyzed separately.

Finally, I’m missing AD occurrence (as a total group) in the prognostic models, given the aim of the study this would also be interesting to evaluate. Why did the authors stratify for early and advanced stage in prognostic modelling? It is also possible to include this variable as a predictive in the model, this would also results in larger groups which increases the power of the analysis.

In conclusion, while the authors have a very interesting dataset, the current analyses do not adequately answer the aims of the study. My suggestion is to remove most of the underpowered subgroup analyses and to focus on a main question that they want to answer, for example occurrence of AD after HCC treatment and impact on prognosis (but then results should be presented separately per treatment and take into account the different HCC populations undergoing these treatments). In the current form it is difficult to place these results in clinical context.

Reply: The authors agree with the points raised by the Reviewer. Following the Reviewer’s suggestion, our study focused on occurrence of AD after HCC treatment and impact on prognosis, and added a comparison between HCC patients with and without cirrhosis. Therefore, we have changed the relevant sentences in the Abstract section (marked-up copy; page 3, lines 5-6, 8-12, 14-16, 18-23, 25; page 4 line 1), the Introduction section (marked-up copy; page 5, lines 21-25; page 6, lines 1, 3-7), the Patients and Methods section (marked-up copy; page 7, lines 5-8, 20-22), the Result section (marked-up copy; page 10, lines 4-5, 7-8, 10-13, 16-24; page 11, lines 2-4, 6-9, 11-25; page 12, lines 1-25; page 13, lines 1-5, 8-9, 14-20; page 14, lines 1-4, 8-25; page 15, lines 1-25; page 16, lines 1-25), the Discussion section (marked-up copy; page 17; lines 3-15; page 18, lines 3-4, 7-9, 17-19, 21-22; page 19, lines 5-8, 9-10, 15, 17-25; page 20, lines 1-3, 8-14), Figure 1, and Tables 2-10.

Minor comments:

- Why did the authors exclude Child-pugh C patients? Or patients with performance status of 2-3?

Reply: Thank you for your comments. According to your major comment, we had changed the aim to clarify the impact of AD after treatment for HCC. Therefore, we changed the exclusion criteria as follows: “Patients receiving maintenance dialysis or who received only best supportive care were excluded.” (marked-up copy; page 7, lines 8-11). In addition, we changed the relative sentences in the Abstract section (marked-up copy; page 3, lines 7, 11, 17), the Results section (marked-up copy; page 10, lines 4, 7, 12-13, 16-18; page 13, lines 8-11, 13-14, 20, 23-24; page 14 line 8), and the Discussion section (marked-up copy; page 18, line 12).

- Was liver biopsy known in every patient? The methods sections suggests this but it is not standard to biopsy every cirrhotic patient

Reply: We apologize for the unclear message in the original manuscript. We modified the relevant sentence (marked-up copy; page 8, line 10).

- The numbers of the ACLF group are too low to perform prognostic modelling.

Reply: According to your comment, we excluded the ACLF group from performing prognostic modeling (marked-up copy; page 14, lines 6-25; page 15, lines 1-12).

Reviewer #2: Kondo et al. performed a retrospective study to evaluate the impact of AD and ACLF on outcome in patients with HCC. Authors were trying to identify independent factors impacting on prognosis in this cohort.

Although this topic is interesting it is a difficult to read paper mostly due to the sequence results are presented. Different types of analysis (frequency of events, prognosis, univariate analysis, multivariate analysis) and distinct subgroups are put together into paragraphs which makes it difficult to comprehend the final messages.

General aspects:

I suggest, and this is my understanding of the manuscript, that authors make sure to address the following aspects and to put them in a more logical order:

• What is the relevance of AD and ACLF in cirrhosis with HCC (frequency and impact on prognosis)

Reply: According to your suggestion, we changed paragraph structure in the Result section as follows: #1 Patient characteristics, #2 Development of AD, #3 Predictive factors for the development of AD stratified with a type of treatment strategy for HCC, #4 Predictive factors for the development of AD stratified with HCC stage, #5 Impact of AD on prognosis, #6 Development of ACLF, #7 Prognostic model for 28/90-day mortality in cases with AD.

• Is HCC and/or the stage of liver disease a factor impacting on the frequency of AD/ACLF (e.g. frequency and impact of HCC treatment related AD/ACLF, comparison with non-treatment related AD/ACLF; how many HCC related and how many AD/ACLF related deaths etc).

Reply: We modified the analyses of the predictive factor of AD development and prognostic factor for short-term mortality after AD diagnosis, taking into account HCC treatment type and HCC stage. Therefore, we changed the relevant sentences in the Results section (marked-up copy; page 11, lines 11-25; page 12, lines 1-25; page 13, lines1-5; page 14, lines 6-25; page 15, lines 1-12) and Tables 2-10. In terms of ACLF, according to Reviewer 1’s suggestion, we excluded the ACLF group to perform prognostic modeling. During the study period, 223 patients died (marked-up copy; page 10, lines 9-10). However, it is difficult to distinguish HCC related death from AD/ACLF related death, especially in patients with advanced-stage HCC.

• Are there further independent factors impacting on patients’ outcome?

Reply: Thank you for your comment. We added Table 9 as analyses for prognostic factors and the sentences in the Result section (marked-up copy; page 12, lines 19-23).

More specific aspects:

I am also missing the justification for dividing the groups upfront into early and late stage HCC. Although it is logical from the oncological point of view this paper is trying to elaborate on the decompensated cirrhosis/ACLF, which is meant to be biologically different. Authors should explain why they chose to do this subgroup analysis. I understand the results as if there are many data suggesting the severity of liver disease rather than the HCC itself as the main determinant for risk of AD/ACLF and subsequent outcome.

Reply: Thank you for your comment. Please also see the response to Reviewer 1’s major comment. We changed the analyses to include HCC stage as a predictive factor and conducted subgroup analyses stratified HCC stage (marked-up copy; Tables 2-10).

Why did authors include patients without cirrhosis? EASL CLIF ACLF definition requires the presence of liver cirrhosis.

Reply: Thank you for your comment. We apologize for the unclear message in the original manuscript. We hypothesize that AD is a crucial event regardless of the presence of cirrhosis. Therefore, we added a comparison between HCC patients with and without cirrhosis to clarify the hypothesis (marked-up copy; page 6, lines 4-7; page 14, lines 1-4). The results showed that AD without cirrhosis had similar ACLF incidence and similar short-term mortality after admission due to AD, compared to AD without cirrhosis.

At baseline, did all patients included have compensated cirrhosis?

Please explain the rational for excluding Child-C patients?

Reply: Thank you for your comment. According to Reviewer 1’s comment, we had changed the aim to clarify the impact of AD after treatment for HCC, so we changed the exclusion criteria as follows: “Patients receiving maintenance dialysis or who received only best supportive care were excluded.” (marked-up copy; page 7, lines 8-11). In addition, we changed the relative sentences in the Abstract section (marked-up copy; page 3, lines 7, 11, 17), the Results section (marked-up copy; page 10, lines 7, 12-13, 16-18; page 13, lines 8-11, 13-14, 20, 23-24; page 14 line 8), and the Discussion section (marked-up copy; page 18, line 12).

How many patients were transplanted during follow up? I understood that eligibility for transplant was no exclusion criterion.

Reply: Thank you for your comment. Only one patient received liver transplantation (marked-up copy; page 10, lines 9-10).

Authors use the terms admission and hospitalisation – is that meant to be a surrogate for admission due to AD/ACLF? It is a bit confusing as admissions in general are not always due to AD/ACLF.

Reply: Thank you for your comment. We changed the relevant sentences (marked-up copy; page3, lines 22-23; page 8, line 9; page 13, line 9; page 14, line 6).

What is meant by curative treatment – intended to be or ultimately curative?

Reply: Thank you for your comment. That intended to be curative, so we changed curative treatment to potentially curative treatment (marked-up copy; page 10, line 8; page 11, line 11; page 12, line 21; Table 9).

How many patients developed AD/ACLF as the consequence of HCC therapy? Are there factors impacting on AD/ACLF after HCC therapy?

Reply: As shown in new Table 10, 35 cases developed AD due to HCC therapy. Three cases developed ACLF as the consequence of HCC therapy. Following your suggestion, we analyzed factors impacting on AD after HCC therapy, but we could not find the significant factor.

Were all patients with viral hepatitis treated with antivirals? Any information on viral load etc? Were there differences between treated and non-treated patients?

Reply: Thank you for your comment. We added some sentences in the Discussion section (marked-up copy; page 18, lines 4-7): 57% of patients with virus-related HCC were treated with antivirals, and there was a significant difference in the development of AD between treated and non-treated patients (treated vs. non-treated: 5.4% vs. 17.6% at 1 year/12.7% vs. 42.4% at 3 years, p < 0.001).

There are too many variables in the regression analysis. For treatment and aetiology it may be useful to have only a binary variable (categorical), e.g. treatment/no treatment or viral hepatitis/others

Reply: Thank you for your suggestion. Following your suggestion, we excluded alcohol and NASH as etiology in the regression analyses and prognostic analyses. In addition, we used potentially curative treatment in the cox regression analyses of predictive factors (marked-up copy; Table 9).

Table 5 – Ascites, HE etc…. are no precipitating events, but complications of cirrhosis defining decompensation.

Reply: We apologize for the inappropriate term and changed “Categories of main precipitating events” to “Complications defining AD” (mar

Submitted filename: Rebuttal letter.docx

Click here for additional data file.

26 Nov 2021

3 Dec 2021

PONE-D-21-23680R1

Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma

PLOS ONE

To the Reviewers

Thank you for your comments and critical review.

Reviewer #1: The article greatly improved by the corrections made by the authors. I have still some minor comments:

- How do the authors explain that cirrhosis is not a risk factor for AD or prognosis when compared to non-cirrhotics?

Reply: Thank you for your comments. In our study, cirrhosis was associated with AD development (marked-up copy; page 9, lines 17-18; page 16, line 11). However, cirrhosis is not a risk factor for long-term mortality. Therefore, we added some sentences for explaining this result in the Discussion section (marked-up copy; page 15, lines 6-11).

- Part of the discussion is on why Albi-score and CLIF-C are associated with prognosis, but if I understand it correctly these were predictors of prognosis in the original study and not in the rebuttal. If this is the case, the authors should focus on discussion of the new predictors if they think this is needed ( ACLF, bilirubin level, HCC progression or MELD score)

Reply: We agree with your suggestion. According to your suggestion, we deleted some sentences in the Discussion section (marked-up copy; page 16, lines 20-25; page 17, line 1).

- They authors suggest to perform surgical resection in patients who have HCC with preserved liver function, even at the intermediate-advanced–stage, because of the low AD incidence rate after treatment. The fact that AD is low in the resection group can also be reflected by the relative good status of patients in this group (younger age, no portal hypertension, few comorbidities). If this is the case, it is possible that the low AD occurence in the surgical group is attributed to the patient group and not to the intervention (selection bias). Did the authors compare baseline characteristics of HCC paitents in the different groups to evaluate this? If these groups are indeed different, the authors should consider to remove this conclusion.

Reply: Thank you for your comments. According to your comments, we compared baseline characteristics between the surgical and non-surgical groups. In fact, patients who received surgery had lower Child-Pugh scores than those who did not, and had no prior history of AD. Therefore, we modified some sentences in the Abstract section (marked-up copy; page 3, lines 14-15, 22-24) and the Discussion section (marked-up copy; page 14, lines 22-23; page 17, lines 10-13).

- Please check the manuscript for spelling errors, the word cirrhosis is spelled incorrectly in some placed (for example methods section abstract)

Reply: We apologize for some errors. We checked and corrected some errors in the Abstract section (marked-up copy; page 3, lines 9, 18), the Patients and Methods section (marked-up copy; page 7, line 23), the Results section (marked-up copy; page 9, line 16; page 12, lines 17-18, 25), and the Discussion section (marked-up copy; page 14, line 21).

Reviewer #2: I would like to thank the authors for their efforts to respond to all our comments. The changes applied to the manuscript improved its quality subtantially.

Authors should go through the manuscript once more to remove some remaining typos.

Reply: We would also like to thank you for taking the time to review our manuscript and apologize for some remaining typos. We checked and corrected the remaining typos in the Abstract section (marked-up copy; page 3, lines 9, 18), the Patients and Methods section (marked-up copy; page 7, line 23), the Results section (marked-up copy; page 9, line 16; page 12, lines 17-18, 25), and the Discussion section (marked-up copy; page 14, line 21).

Submitted filename: Rebuttal letter.docx

Click here for additional data file.

7 Dec 2021

Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma

PONE-D-21-23680R2

Dear Dr. Kondo,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Pavel Strnad

Academic Editor

PLOS ONE

Additional Editor Comments (optional): Thank you for submitting you interesting work to PLoS One!

Reviewers' comments:

19 Jan 2022

PONE-D-21-23680R2

Impact of acute decompensation on the prognosis of patients with hepatocellular carcinoma

Dear Dr. Kondo:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Pavel Strnad

Academic Editor

PLOS ONE

Table 4

Cox regression analyses of predictive factors for the development of acute decompensation in the ablation group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	1.004 (0.972–1.038)	0.805	-
Male sex	0.941 (0.556–1.595)	0.822	-
Liver cirrhosis	3.535 (1.514–8.252)	0.004	-
Prior history of acute decompensation	3.443 (1.852–6.401)	<0.001	2.182 (1.143–4.166)	0.018
Ascites	3.083 (1.455–6.530)	0.003	-
Intermediate-advanced-stage HCC	4.362 (1.334–14.264)	0.015	4.418 (1.295–15.075)	0.018
Etiology
Virus related hepatitis	0.438 (0.255–0.754)	0.029	0.445 (0.249–0.795)	0.006
Laboratory data
Alanine aminotransferase (U/L)	0.999 (0.991–1.006)	0.737	-
Bilirubin (mg/dL)	1.957 (1.111–3.450)	0.021	-
Prothrombin time (international normalized ratio)	4.128 (0.381–44.691)	0.243	-
Albumin (g/dL)	0.343 (0.190–0.618)	<0.001	-
Creatinine (mg/dL)	1.253 (0.471–3.331)	0.651	-
Sodium (mmol/L)	0.976 (0.907–1.051)	0.522	-
Platelets (109/L)	0.934 (0.883–0.989)	0.018	-
Alfa fetoprotein > 18 ng/ml*	1.838 (1.097–3.080)	0.021	2.020 (1.174–3.475)	0.011
Albumin-bilirubin (ALBI) score	3.354 (1.810–6.215)	<0.001	2.606 (1.386–4.903)	0.003
Child-Pugh score	1.812 (1.327–2.494)	<0.001	-
Model for end-stage liver disease (MELD) score	1.065 (0.855–1.328)	0.573	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.

Table 5

Cox regression analyses of predictive factors for the development of acute decompensation in the chemoembolization group.

	Univariate hazard ratio (95% confidence interval)	P value	Multivariate hazard ratio (95% confidence interval)	P value
Age	0.980 (0.958–1.001)	0.067	-
Male sex	1.146 (0.680–1.933)	0.609	-
Liver cirrhosis	1.414 (0.854–2.341)	0.179	-
Prior history of acute decompensation	2.301 (1.431–3701)	0.001	-
Ascites	1.622 (0.944–2.787)	0.080	-
Intermediate-advanced-stage HCC	2.260 (1.360–3.756)	0.002	2.359 (1.405–3.962)	0.001
Etiology
Virus related hepatitis	0.805 (0.523–1.238)	0.323	-
Laboratory data
Alanine aminotransferase (U/L)	1.005 (1.002–1.008)	0.004	1.005 (1.001–1.009)	0.009
Bilirubin (mg/dL)	2.451 (1.794–3.350)	<0.001	1.594 (1.097–2.315)	0.014
Prothrombin time (international normalized ratio)	6.725 (2.155–20.990)	0.001	-
Albumin (g/dL)	0.358 (0.235–0.546)	<0.001	-
Creatinine (mg/dL)	1.059 (0.582–1.928)	0.852	-
Sodium (mmol/L)	0.929 (0.873–0.988)	0.019	-
Platelets (109/L)	1.005 (0.983–1.028)	0.638	-
Alfa fetoprotein > 18 ng/ml*	1.250 (0.809–1.933)	0.314	-
Albumin-bilirubin (ALBI) score	3.266 (2.113–5.048)	<0.001	-
Child-Pugh score	1.819 (1.506–2.197)	<0.001	1.681 (1.323–2.135)	<0.001
Model for end-stage liver disease (MELD) score	1.097 (1.015–1.187)	0.020	-

HCC; hepatocellular carcinoma.

*Cut-off value was determined by receiver operating characteristics analysis.