Wang X, Lin ZY, Zhou Y, Zhong Q, Li ZR, Lin XX, Hu MG, He KL. Association of preoperative antiviral treatment with incidences of post-hepatectomy liver failure in hepatitis B virus-related hepatocellular carcinoma. World J Gastrointest Surg 2024; 16(7): 2106-2118 [PMID: 39087126 DOI: 10.4240/wjgs.v16.i7.2106]
Corresponding Author of This Article
Kun-Lun He, MD, Chief Physician, Professor, Medical Big Data Research Center, Chinese PLA General Hospital, No. 28 Fuxing Road, Haidian District, Beijing 100853, China. hekunlun301dr@163.com
Research Domain of This Article
Gastroenterology & Hepatology
Article-Type of This Article
Retrospective Study
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
World J Gastrointest Surg. Jul 27, 2024; 16(7): 2106-2118 Published online Jul 27, 2024. doi: 10.4240/wjgs.v16.i7.2106
Association of preoperative antiviral treatment with incidences of post-hepatectomy liver failure in hepatitis B virus-related hepatocellular carcinoma
Xiao Wang, You Zhou, Qin Zhong, Zong-Ren Li, Xi-Xiang Lin, Kun-Lun He, Medical Big Data Research Center, Chinese PLA General Hospital, Beijing 100853, China
Xiao Wang, Zhao-Yi Lin, Qin Zhong, Zong-Ren Li, Xi-Xiang Lin, Medical School of Chinese PLA, Chinese PLA General Hospital, Beijing 100853, China
Xiao Wang, Department of Hepatobiliary Surgery, Chinese PLA 970th Hospital, Yantai 264001, Shandong Province, China
Zhao-Yi Lin, Ming-Gen Hu, Faculty of Hepato-Pancreato-Biliary Surgery, The First Medical Center of Chinese PLA General Hospital, Beijing 100853, China
You Zhou, School of Medicine, Nankai University, Tianjin 300071, China
Co-corresponding authors: Ming-Gen Hu and Kun-Lun He.
Author contributions: Wang X, Lin ZY, Hu MG and He KL conceptualized and designed the research study; Wang X, Lin ZY, Zhong Q, Li ZR and Lin XX performed data collection and collation; Wang X, Lin ZY and Zhou Y completed data analysis; Wang X and Lin ZY contributed to the writing of the original manuscript; Hu MG and He KL supervised the research process and revised the manuscript; All authors have read and approved the final submitted version of the manuscript. Wang X proposed and designed the research approach, performed data analysis and prepared the first draft of the manuscript. Lin ZY was mainly responsible for patient screening and collection of clinical data. Both authors have made crucial contributions towards the completion of the project and thus qualified as the co-first authors of the paper. Both He KL and Hu MG have played indispensable roles in the experimental design, data interpretation and manuscript preparation as the co-corresponding authors. He KL applied for and obtained the funds for this research project. He KL conceptualized, designed, and supervised the whole process of the project. He searched the literature, revised and submitted the early version of the manuscript with the focus on the association between antiviral therapy and post-hepatectomy liver failure. Hu MG was instrumental for comprehensive literature search, preparation and submission of the current version of the manuscript with a focus on hepatitis B virus DNA level as a predictor of post-hepatectomy liver failure. This collaboration between He KL and Hu MG is crucial for the publication of this manuscript.
Supported byScience and Technology Innovation 2030 - Major Project, No. 2021ZD0140406 and No. 2021ZD0140401.
Institutional review board statement: This study was approved by the Medical Ethics Committee of Chinese PLA General Hospital (No. S2022-766-01).
Informed consent statement: The informed consent has been waived by the Medical Ethics Committee of Chinese PLA General Hospital.
Conflict-of-interest statement: All the authors declare no competing interests for this article.
Data sharing statement: All data generated or analyzed during this study were included in this published article. The informed consent was not obtained but the presented data are anonymized and risk of identification is low. The original clinical data used in this study can be available with the approval of the corresponding author at hekunlun301dr@163.com.
Open-Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Kun-Lun He, MD, Chief Physician, Professor, Medical Big Data Research Center, Chinese PLA General Hospital, No. 28 Fuxing Road, Haidian District, Beijing 100853, China. hekunlun301dr@163.com
Received: March 12, 2024 Revised: June 3, 2024 Accepted: June 20, 2024 Published online: July 27, 2024 Processing time: 132 Days and 1.9 Hours
Abstract
BACKGROUND
Post-hepatectomy liver failure (PHLF) is a common consequence of radical partial hepatectomy in hepatocellular carcinoma (HCC).
AIMS
To investigate the relationship between preoperative antiviral therapy and PHLF, as well as assess the potential efficacy of hepatitis B virus (HBV) DNA level in predicting PHLF.
METHODS
A retrospective study was performed involving 1301 HCC patients with HBV who underwent radical hepatectomy. Receiver operating characteristic (ROC) analysis was used to assess the capacity of HBV DNA to predict PHLF and establish the optimal cutoff value for subsequent analyses. Logistic regression analyses were performed to assess the independent risk factors of PHLF. The increase in the area under the ROC curve, categorical net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were used to quantify the efficacy of HBV DNA level for predicting PHLF. The P < 0.05 was considered statistically significant.
RESULTS
Logistic regression analyses showed that preoperative antiviral therapy was independently associated with a reduced risk of PHLF (P < 0.05). HBV DNA level with an optimal cutoff value of 269 IU/mL (P < 0.001) was an independent risk factor of PHLF. All the reference models by adding the variable of HBV DNA level had an improvement in area under the curve, categorical NRI, and IDI, particularly for the fibrosis-4 model, with values of 0.729 (95%CI: 0.705-0.754), 1.382 (95%CI: 1.341-1.423), and 0.112 (95%CI: 0.110-0.114), respectively. All the above findings were statistically significant.
CONCLUSION
In summary, preoperative antiviral treatment can reduce the incidence of PHLF, whereas an increased preoperative HBV DNA level has a correlative relationship with an increased susceptibility to PHLF.
Core Tip: The correlation between preoperative antiviral treatment and post-hepatectomy liver failure (PHLF), hepatitis B virus (HBV) DNA level, and PHLF were analyzed through univariate and multivariate logistic regression analyses, respectively. The optimal cut-off value of HBV DNA was determined through Receiver operating characteristic (ROC) curve analysis. Based on three conventional scoring models, the efficacy of HBV DNA level in predicting PHLF was quantified by computing the area under the ROC curve, categorical net reclassification improvement, and integrated discrimination improvement.
Citation: Wang X, Lin ZY, Zhou Y, Zhong Q, Li ZR, Lin XX, Hu MG, He KL. Association of preoperative antiviral treatment with incidences of post-hepatectomy liver failure in hepatitis B virus-related hepatocellular carcinoma. World J Gastrointest Surg 2024; 16(7): 2106-2118
Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer[1], with its incidence majorly attributed to hepatitis B virus (HBV) infection[2], specifically among Asian male populations[3]. With the advancement of medical technology and the limitations of other treatment options[4,5], hepatectomy remains the major curative intervention for resectable HCC patients[1,6]. Despite the preoperative evaluation and screening of patients, post-hepatectomy liver failure (PHLF) remains the most serious complication with a potential risk of mortality in clinical practice[7,8]. Identifying risk factors associated with PHLF is critical to improving the prognosis of HCC resection.
PHLF refers to a clinical syndrome originating from surgical shock or reduced hepatic parenchymal volume, its occurrence hinging upon two pivotal determinants, i.e., the residual quality and quantity of the liver[9]. In addition to conventional clinical models such as Child-Pugh stage[10] and model for end-stage liver disease scoring system[11], which are employed for PHLF prediction, several improved scoring models, such as the albumin-bilirubin (ALBI) score[12], fibrosis-4 (FIB-4) score[13], and aspartate aminotransferase-to-platelet ratio index (APRI)[14], have been identified as autonomous prognosticators of PHLF. Furthermore, the infiltration of inflammatory cells and factors within the tumor microenvironment plays an important role in tumor progression and the development of cirrhosis, both of which are closely associated with poor adverse prognoses[15-18]. Notably, HBV infection causes a widespread inflammatory response by stimulating the immune system and producing a vast array of inflammatory mediators[19]. Consequently, antiviral treatment (AVT) has been reported to ameliorate the hepatic inflammatory microenvironment by inhibiting viral replication[20,21].
To our knowledge, the level of HBV DNA acts as a direct, specific, and sensitive indicator for HBV infection. It constitutes an important reference for diagnosis, treatment, and assessment of the condition in patients with HBV-related ailments[22]. Previous studies have underscored the common occurrence of HBV reactivation in patients diagnosed with HBV-related HCC undergoing partial hepatectomy[23]. Prophylactic AVT has been shown to block the replication of HBV DNA, mitigate liver inflammation, and forestall HBV reactivation[24]. These findings point to a likely relationship between preoperative AVT and PHLF-associated HBV DNA levels.
This study sought to explore the relationship between preoperative AVT and the occurrence of PHLF in addition to evaluating the value of HBV DNA levels in predicting PHLF.
MATERIALS AND METHODS
Patients
In this retrospective study, patients who underwent radical hepatectomy at the First Medical Center of the General Hospital of the People's Liberation Army were recruited between January 2012 and December 2021. Thorough data desensitization measures were performed to protect the personal information of study participants. Ethical approval for the study was obtained from the Ethics Committee of the First Medical Center of the General Hospital of the People's Liberation Army in adherence to the Declaration of Helsinki. Figure 1 shows the flowchart of the study design.
Figure 1 Flow chart of the patient enrolling process.
HCC: Hepatocellular carcinoma; HBV: Hepatitis B virus; HBsAg: Hepatitis B surface antigen.
Inclusion and exclusion criteria
Eligibility criteria included: (1) Patients with HCC confirmed by pathological examination; (2) Patients who had undergone relevant imaging [enhanced computed tomography (CT) or magnetic resonance imaging (MRI)] within 1 month before surgery; (3) No history of tumor treatment before hepatectomy; (4) No preoperative obstructive jaundice; and (5) No coexisting malignancies or distant metastasis.
The exclusion criteria included: (1) Patients without HBV infection; (2) Concurrent hepatitis C virus antibody positivity; (3) Patients whose HBV DNA test was not performed before surgery; or (4) No preoperative medical record data.
Clinical variables
The determination and value of all variables were extracted from electronic medical records and the outcomes from preoperative assessments. Demographic characteristics included age, sex, a history of diabetes and hypertension, body mass index, hepatitis B, and hepatitis C status, obtained from admission records. Surgical information including operation type, the extent of hepatectomy, volume of blood loss, and the necessity for blood transfusion were obtained from surgical records. Routine preoperative investigations included serological assays and medical imaging. Preoperative serum analysis incorporated metrics such as white blood cell count, hemoglobin, platelet count (PLT), aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin (Alb), gamma-glutamyl transpeptidase (γ-GT), prothrombin time (PT), fibrinogen (FIB), alkaline phosphatase (ALP), total bilirubin (TBil), blood urea nitrogen, creatinine (Cr), international normalized ratio (INR), alpha-fetoprotein (AFP), hepatitis B e antigen (HBeAg), and HBV DNA.
Moreover, the imaging data of patients provided tumor-related features including tumor size (major nodule diameter), tumor number, and determination of cirrhosis. These data were initially extracted from enhanced MRI reports and confirmed from enhanced CT scans. Additionally, the presence of clinically significant portal hypertension (CSPH) was defined by observing esophageal varicosity detected during endoscopy, a history of esophageal variceal bleeding, or substantial splenomegaly (major diameter > 12 cm) coupled with a PLT count < 100 × 109/L[25]. The assignment of Child-Pugh grade primarily relied on three objective variables (TBil, Alb, PT), and two subjective variables (ascites, hepatic encephalopathy)[26]. Evaluation of these parameters was carried out by two proficient and independent clinicians (Xiao Wang and Zhao-Yi Lin). Disagreements arising were resolved through deliberation until a consensus was reached.
Preoperative AVT and patient grouping
Based on the previously established definition, we referred to preoperative AVT as the continuous use of antiviral drugs for more than 90 days before surgery, without restrictions on the type of drugs. Within our study cohort, non-compliance criteria for preoperative AVT included no treatment, previous treatment, discontinuation of treatment, and short-term treatment.
ROC analysis was used as an effective threshold selection method to assess the capacity of HBV DNA values to predict PHLF[27,28]. Among the cutoff values with adequate sensitivity, the value with the highest Youden’s index was selected as the optimal cutoff value for subsequent analyses. Based on the optimal cutoff value obtained, the results of HBV DNA were divided into high-level and low-level groups before conducting subsequent data analysis.
All patients with HBV-related HCC received oral antiviral drugs including entecavir or tenofovir disoproxil after surgery.
Liver resection
All the recruited patients underwent radical surgical resection[29], using various surgical approaches including open surgery, laparoscopy, or robotic surgery. A curative hepatectomy was defined as a surgical procedure designed for complete nodule removal, ensuring that microscopic tumor margins were devoid of malignancy. Specific surgical techniques for partial hepatectomy included local tumor resection, segmental hepatectomy, lobectomy, hemihepatectomy, and extended hemihepatectomy. Intraoperatively, the Pringle maneuver was used to occlude hepatic blood flow when deemed necessary, hence promoting a smooth surgical procedure and mitigating intraoperative hemorrhage.
Definition of PHLF
In light of data retrieval difficulties and clinical practicalities, there were instances where certain patients did not concurrently undergo reviews for both TBil and INR on or after postoperative day 5. To mitigate potential analysis bias from the exclusion of the dataset, we adopted a composite approach that combined two international standards, including the International Study Group of Liver Surgery (ISGLS)[30] and Memorial Sloan Kettering Cancer Centre (MSKCC) criteria[31] to delineate the occurrence of PHLF.
Specific criteria for PHLF identification, applicable on or after postoperative day 5, encompassed the following conditions: A value of TBil > 24 μmol/L and INR > 1.2, or TBil > 70.1 μmol/L (4.1 mg/dL), or INR > 2.5, or ascites drainage > 500 mL/day. Furthermore, patients who were discharged from the hospital within 5 days after surgery were categorized as not experiencing liver failure, except for those discharged due to bleeding or those who had succumbed.
Reference score prediction models for PHLF
In our comparative analysis, we used three prevalent scoring models as reference models for predicting PHLF. The respective calculation formulas for these models are as follows:
ALBI score was calculated as described: [log10 bilirubin (μmol/L) × 0.66] + [albumin (g/L) × (-0.085)].
FIB-4 score, as a simple index, was developed: [age (year) × AST (U/L)]/[PLT count (109/L) × ALT (U/L)1/2].
APRI score was calculated as described: [AST (U/L)/upper limit of normal (U/L)]/PLT count (109/L) × 100.
Statistical analysis
All categorical variables were expressed as the corresponding percentage and compared using the χ2 test. All continuous variables were expressed as means and SD or as medians and interquartile ranges (IQR) and compared using the unpaired t-test (normal distribution) or Mann-Whitney U test (abnormal distribution). The incidence of missing data varied across individual variables, ranging from 0.08% for blood loss to 8.53% for HBeAg, as illustrated in the Supplementary Table 1. To obviate the necessity for case deletion in subsequent analyses, a multiple imputation method was used to address the missing data. Preoperative factors with potential relationships with the incidence of PHLF were assessed through univariate and multivariable logistic regression analyses. Subsequently, upon incorporating meaningful factors into the existing scoring model, we assessed the change in the area under the ROC curve (ΔAUC) between the two models. The statistical significance of area under the curve (AUC) differences was ascertained using Delong's test. Additionally, Decision Curve Analysis (DCA) curves were constructed to evaluate the utility of the model, facilitated by the rmda package. Furthermore, we computed the categorical net reclassification improvement (NRI) and integrated discrimination improvement (IDI)[32-35] from the inclusion of the HBV DNA level, leveraging the predictABEL package. Statistical significance was denoted by a threshold of P < 0.05. All statistical analyses were conducted using the R software (version 4.2.1).
RESULTS
Patient characteristics
This study enrolled 1301 patients (1114 men, 85.6%) who had undergone radical partial hepatectomy for HBV-related HCC, among which 168 (12.9%) had PHLF events. Furthermore, 360 patients (26.8%) met the preoperative AVT criteria and 850 patients (65.3%) did not receive any preoperative AVT. Among the remaining 91 patients (7.0%) who did not meet preoperative AVT criteria, 8 patients had discontinuous treatment, 53 patients stopped taking medication due to improvement in previous AVT, and 30 patients received preoperative short-term AVT with a median (IQR) treatment duration of 15.5 (10-30) days before surgery. Among all patients who were treated with AVT, nucleos(t)ide analogues were the most commonly administered, such as lamivudine, entecavir, and tenofovir.
Among the patients who met our preoperative AVT criteria (n = 360), PHLF occurred in 29 patients (8.0%). Out of the patients who did not receive preoperative AVT (n = 941), PHLF occurred in 139 patients (14.7%). Moreover, patients who did not meet the preoperative AVT criteria had increased liver function indicators (ALT, AST, γ-GT, Alb, ALP), increased levels of AFP and HBV DNA, larger tumor diameter, greater intraoperative blood loss, and an increased likelihood of needing blood transfusions. However, we found no statistically significant differences between the two groups regarding gender, HBeAg, cirrhosis, tumor number, CSPH, or comorbidities such as diabetes and hypertension. Table 1 presents comprehensive statistical data for all clinical variables.
Table 1 Basic characteristics in hepatitis B virus-related hepatocellular carcinoma patients (antiviral treatment and non-antiviral treatment group), n (%).
The capacity of HBV DNA values to predict PHLF was analyzed using the ROC curves (Figure 2). We eventually defined the optimal critical value as 269 IU/mL, with an AUC of 0.642 (95%CI: 0.615-0.668; P < 0.001; Table 2). Among the 941 patients who did not receive AVT, 605 individuals had a high viral load, as determined by the HBV DNA level exceeding the optimal critical value of 269 IU/mL. In this subset, the incidence of PHLF was substantially higher at 19.5%, compared to the group with a low viral load where the incidence was 6.2%. Additionally, statistically significant variations were observed in ALT, AST, γ-GT, Alb, ALP, FIB, PT, INR, AFP, HBeAg, and tumor size. Table 1 presents comprehensive data statistics for all clinical variables.
Figure 2 The receiver operating characteristic curve of hepatitis B virus DNA value for predicting the occurrence of post-hepatectomy liver failure.
Table 2 The area under the receiver operating characteristic curve and its cut-off value for hepatitis B virus DNA value according to an increase of post-hepatectomy liver failure in hepatitis B virus-related hepatocellular carcinoma patients.
AUC (95%CI)
HBV DNA value (IU/mL)
Sensitivity
1-Specificity
P value
PHLF
0.642 (0.615-0.668)
269
0.75
0.522
< 0.001
The relationship between preoperative AVT and PHLF in HBV-related HCC patients
In the study cohort, the univariate analysis revealed that the presence or absence of AVT before surgery significantly influenced the incidence of PHLF (P = 0.001); on the other hand, multivariate analysis showed no statistical significance (P = 0.624; Table 3). Subsequently, in a multivariate analysis with the variable of HBV DNA level excluded, preoperative AVT independently correlated with a reduced risk of PHLF [odds ratio (OR) 0.61; 95%CI: 0.38-0.98] (P = 0.039; Table 4).
Table 3 Logistic regression analysis of the post-hepatectomy liver failure in hepatitis B virus-related hepatocellular carcinoma patients (n = 1301).
Variable
Univariate
Multivariate
OR (95%CI)
P Value
OR (95%CI)
P Value
Sex, male vs female
1.83 (1.05-3.18)
0.033
1.62 (0.80-3.30)
0.181
Age, year (continuous)
1.02 (1.00-1.03)
0.061
1.04 (1.01-1.06)
0.002
Hypertension, yes vs no
0.73 (0.49-1.09)
0.118
0.88 (0.53-1.47)
0.579
Diabetes, yes vs no
0.73 (0.48-1.13)
0.156
0.75 (0.45-1.28)
0.292
BMI, kg/m2 (continuous)
0.97 (0.93-1.03)
0.319
0.97 (0.91-1.03)
0.296
WBC count, 109/L (continuous)
0.89 (0.81-0.99)
0.024
0.95 (0.83-1.08)
0.398
Hemoglobin, g/L (continuous)
1.01 (1.00-1.02)
0.249
1.02 (1.00-1.03)
0.013
PLT, 109/L (continuous)
0.99 (0.99-1.00)
< 0.001
0.99 (0.99-1.00)
0.001
ALT, U/L (continuous)
1.01 (1.00-1.01)
0.001
1.00 (0.99-1.00)
0.534
AST, U/L (continuous)
1.02 (1.01-1.02)
< 0.001
1.01 (1.00-1.02)
0.261
γ-GT, U/L (continuous)
1.00 (1.00-1.01)
< 0.001
1.00 (1.00-1.00)
0.046
Alb, g/L (continuous)
0.88 (0.84-0.92)
< 0.001
0.94 (0.89-0.99)
0.024
ALP, U/L (continuous)
1.00 (1.00-1.01)
0.003
1.00 (1.00-1.00)
0.999
TBil, μmol/L (continuous)
1.08 (1.06-1.11)
< 0.001
1.07 (1.04-1.10)
< 0.001
BUN, mmol/L (continuous)
0.89 (0.78-1.00)
0.059
0.88 (0.76-1.03)
0.123
Cr, μmol/L (continuous)
1.00 (0.98-1.01)
0.536
1.00 (0.99-1.02)
0.634
FIB, g/L (continuous)
0.93 (0.82-1.05)
0.223
0.95 (0.82-1.10)
0.476
PT, second (continuous)
1.57 (1.33-1.86)
< 0.001
1.03 (0.66-1.60)
0.913
INR (continuous)
196.32 (32.04-1202.99)
< 0.001
5.45 (0.04-755.25)
0.501
AFP, > 400 vs ≤ 400 ng/mL
1.77 (1.27-2.47)
0.001
1.16 (0.77-1.78)
0.477
HBeAg positivity, yes vs no
1.03 (0.71-1.48)
0.881
0.85 (0.55-1.32)
0.475
Cirrhosis, yes vs no
1.76 (1.27-2.44)
0.001
1.66 (1.10-2.49)
0.015
No. of tumor > 1, yes vs no
0.89 (0.59-1.35)
0.595
0.71 (0.45-1.14)
0.160
Tumor diameter, mm (continuous)
1.01 (1.01-1.02)
< 0.001
1.01 (1.00-1.02)
0.020
Open surgery, yes vs no
0.73 (0.58-0.91)
0.006
0.93 (0.71-1.20)
0.555
Major hepatectomy, yes vs no
2.95 (2.10-4.16)
< 0.001
2.36 (1.48-3.76)
< 0.001
Blood loss > 400 mL, yes vs no
2.20 (1.56-3.09)
< 0.001
1.28 (0.80-2.05)
0.303
Transfusion, yes vs no
2.17 (1.40-3.39)
0.001
1.04 (0.56-1.91)
0.913
CSPH, yes vs no
2.28 (1.48-3.52)
< 0.001
1.38 (0.75-2.52)
0.3
Child-Pugh grade, A vs B
4.88 (1.83-13.01)
0.002
0.88 (0.22-3.48)
0.856
AVT, yes vs no
0.51 (0.33-0.77)
0.001
0.88 (0.53-1.47)
0.624
HBV DNA level, high vs low
3.27 (2.26-4.73)
< 0.001
2.08 (1.10-3.93)
< 0.001
Table 4 Multivariate logistic regression analyses for the incidence of post-hepatectomy liver failure in the subsample of hepatitis B virus-related hepatocellular carcinoma patients.
Variable
Overall patients (n = 1301)
Not receiving AVT (n = 941)
OR (95%CI)
P value
OR (95%CI)
P value
Sex, male vs female
1.48 (0.74-2.99)
0.27
1.37 (0.63-2.99)
0.43
Age, year (continuous)
1.04 (1.01-1.06)
0.002
1.03 (1.00-1.05)
0.046
Hypertension, yes vs no
0.90 (0.56-1.45)
0.667
0.88 (0.51-1.50)
0.63
Diabetes, yes vs no
0.74 (0.44-1.24)
0.245
0.73 (0.40-1.32)
0.296
BMI, kg/m2 (continuous)
0.96 (0.91-1.02)
0.228
0.96 (0.90-1.03)
0.304
WBC count, 109/L (continuous)
0.93 (0.82-1.05)
0.241
0.95 (0.83-1.09)
0.465
Hemoglobin, g/L (continuous)
1.02 (1.01-1.03)
0.007
1.01 (1.00-1.03)
0.067
PLT, 109/L (continuous)
0.99 (0.99-1.00)
0.001
0.99 (0.99-1.00)
0.003
ALT, U/L (continuous)
1.00 (0.99-1.01)
0.939
0.99 (0.98-1.00)
0.179
AST, U/L (continuous)
1.01 (1.00-1.01)
0.328
1.01 (1.00-1.02)
0.127
γ-GT, U/L (continuous)
1.00 (1.00-1.00)
0.030
1.00 (1.00-1.00)
0.103
Alb, g/L (continuous)
0.92 (0.87-0.98)
0.005
0.92 (0.86-0.98)
0.007
ALP, U/L (continuous)
1.00 (1.00-1.00)
0.852
1.00 (1.00-1.01)
0.963
TBil, μmol/L (continuous)
1.07 (1.03-1.10)
< 0.001
1.07 (1.04-1.11)
< 0.001
BUN, mmol/L (continuous)
0.89 (0.76-1.04)
0.129
0.94 (0.79-1.12)
0.48
Cr, μmol/L (continuous)
1.00 (0.99-1.02)
0.578
1.00 (0.98-1.02)
0.853
FIB, g/L (continuous)
0.95 (0.82-1.10)
0.463
0.96 (0.82-1.13)
0.624
PT, second (continuous)
1.00 (0.65-1.56)
0.991
1.08 (0.65-1.79)
0.778
INR (continuous)
8.82 (0.07-1159.35)
0.382
2.91 (0.01-820.09)
0.711
AFP, > 400 vs ≤ 400 ng/mL
1.23 (0.82-1.86)
0.318
1.19 (0.75-1.88)
0.459
HBeAg positivity, yes vs no
0.98 (0.64-1.50)
0.920
1.00 (0.62-1.61)
0.999
Cirrhosis, yes vs no
1.70 (1.13-2.55)
0.011
1.70 (1.08-2.69)
0.022
No. of tumor > 1, yes vs no
0.72 (0.45-1.15)
0.174
0.82 (0.49-1.37)
0.452
Tumor diameter, mm (continuous)
1.01 (1.00-1.02)
0.010
1.01 (1.00-1.01)
0.286
Open surgery, yes vs no
0.91 (0.70-1.17)
0.445
0.98 (0.74-1.32)
0.911
Major hepatectomy, yes vs no
2.71 (1.72-4.29)
< 0.001
2.94 (1.76-4.90)
< 0.001
Blood loss > 400 mL, yes vs no
1.25 (0.78-2.00)
0.347
1.48 (0.88-2.48)
0.141
Transfusion, yes vs no
1.04 (0.56-1.90)
0.912
1.02 (0.52-1.98)
0.963
CSPH, yes vs no
1.28 (0.70-2.34)
0.416
1.21 (0.59-2.50)
0.608
Child-Pugh grade, A vs B
0.92 (0.24-3.50)
0.898
0.54 (0.11-2.58)
0.437
AVT, yes vs no
0.61 (0.38-0.98)
0.039
-
-
HBV DNA level, high vs low
-
2.61 (1.50-4.53)
0.001
The relationship between HBV DNA level and PHLF in HBV-related HCC patients
Univariate and multivariate analyses revealed that the HBV DNA level significantly influenced the incidence of PHLF (P < 0.001; Table 3). To further investigate the relationship between HBV DNA level and the incidence of PHLF, while mitigating any potential confounding effects of AVT, multivariate analyses were exclusively conducted on the cohort of 941 patients who did not receive AVT. The results revealed that HBV DNA high levels (≥ 269 IU/mL) emerged as an independent risk factor, significantly increasing the risk of PHLF (OR 2.61; 95%CI: 1.50-4.53) (P = 0.001). This was consistent even when accounting for other influential factors including PLT, Alb, TBil, cirrhosis, and major hepatectomy (Table 4).
Univariate and multivariate logistic regression analysis for the incidence of PHLF
As shown in Table 3, the majority of variables had a P value below 0.05 in univariate regression analysis. Upon subjecting these variables to multivariate logistic analysis however, Age (P = 0.002), Hemoglobin (P = 0.013), PLT (P = 0.001), γ-GT (P = 0.046), Alb (P = 0.024), TBil (P < 0.001), cirrhosis (P = 0.015), tumor diameter (P = 0.020), major hepatectomy (P < 0.001) and the HBV DNA level (high vs low) (P < 0.001) were the independent risk factors linked to PHLF. In contrast, the multivariate analysis encompassing all variables across the entire patient cohort did not reveal statistical significance for preoperative AVT in relation to PHLF.
Assessment of the incremental value of HBV DNA level over reference scoring models
The scoring results were computed from data extracted from the study cohort and each model showed acceptable predictive performance, with an AUC of 0.672 (95%CI: 0.647-0.698) for ALBI, 0.675 (95%CI: 0.650-0.701) for FIB-4, and 0.695 (95%CI: 0.670-0.720) for APRI. All models showed an improvement in AUC after incorporating variable HBV DNA levels with significant statistical differences (P < 0.05), including ALBI (ΔAUC = 0.037) (P = 0.011), FIB-4 (ΔAUC = 0.054) (P = 0.001), and APRI (ΔAUC = 0.030) (P = 0.003). Figure 3 shows the ROC curves of each scoring model for predicting PHLF. Figure 4 illustrates the DCA curves, used to validate the clinical utility of each model.
Figure 3 Comparison of the receiver operating characteristic curves for predicting post-hepatectomy liver failure using each reference score model alone and the new model combining hepatitis B virus DNA level.
A: The receiver operating characteristic (ROC) curves of albumin-bilirubin (ALBI) and ALBI + hepatitis B virus (HBV) DNA level; B: The ROC curves of fibrosis-4 (FIB-4) and FIB-4 + HBV DNA level; C: The ROC curves of aspartate aminotransferase-to-platelet ratio index (APRI) and APRI + HBV DNA level. ROC: Receiver operating characteristic; ALBI: Albumin-bilirubin; AUROC: The area under the receiver operating characteristic; HBV: Hepatitis B virus; FIB-4: Fibrosis-4.
Figure 4 Decision curve of the reference score models and the developed models combining hepatitis B virus DNA level.
A: Decision curves of albumin-bilirubin (ALBI) and ALBI + hepatitis B virus (HBV) DNA level; B: Decision curves of fibrosis-4 (FIB4) and FIB4 + HBV DNA level; C: Decision curves of aspartate aminotransferase-to-platelet ratio index (APRI) and APRI + HBV DNA level. ALBI: Albumin-bilirubin; HBV: Hepatitis B virus; FIB-4: Fibrosis-4.
Furthermore, each score model combined with HBV DNA level had a positive categorical NRI, with a value of 0.940 (95%CI: 0.886-0.994) for ALBI, 1.382 (95%CI: 1.341-1.423) for FIB-4, and 1.374 (95%CI: 1.330-1.416) for APRI. For IDI, FIB-4, when augmented with HBV DNA level, demonstrated the highest value of 0.112 (95%CI: 0.110-0.114), whereas the values of the other two models were 0.082 (95%CI: 0.080-0.084) and 0.095 (95%CI: 0.094-0.097), respectively. Importantly, all the above statistical results had extremely significant differences (P < 0.0001). Table 5 presents the comprehensive results of all these measures used to assess the incremental value of HBV DNA level in predicting PHLF.
Table 5 Evaluation of the incremental value of hepatitis B virus DNA level in predicting post-hepatectomy liver failure.
Model
AUC (95%CI)
ΔAUC
P value
Categorical NRI (95%CI)
P value
IDI (95%CI)
P value
ALBI + HBV DNA
0.709 (0.684-0.734)
0.037
0.011
0.940 (0.886-0.994)
< 0.0001
0.082 (0.080-0.084)
< 0.0001
FIB4 + HBV DNA
0.729 (0.705-0.754)
0.054
0.001
1.382 (1.341-1.423)
< 0.0001
0.112 (0.110-0.114)
< 0.0001
APRI + HBV DNA
0.725 (0.701-0.749)
0.030
0.003
1.374 (1.330-1.416)
< 0.0001
0.095 (0.094-0.097)
< 0.0001
DISCUSSION
HBV infection is the leading cause of liver injury, cirrhosis, and HCC, in China. Although the HBV DNA level can indicate the extent of virus replication and infectivity, it does not directly represent the extent of liver function impairment or disease severity. Preoperative AVT improves the liver microenvironment and function[20,21,24], with HBV DNA level acting as the major clinical reference for AVT. Despite the strides made in medical technology, PHLF however remains a significant and potentially deadly condition after hepatectomy in patients with HBV-related HCC[7]. Previous studies have investigated the relationship between preoperative AVT or HBV DNA level and the incidence of microvascular invasion in patients with HBV-related HCC[36]. The findings have shown a potential correlation between preoperative AVT or HBV DNA level and PHLF in a cohort of HBV-related HCC patients receiving surgical resection as initial treatment.
Our findings revealed that the incidence of PHLF is significantly lower in patients who received preoperative AVT than those who did not. Additionally, among patients who did not undergo AVT, those with a high-level HBV DNA exhibited a significantly higher incidence of PHLF than those with a low-level HBV DNA. Patients with non-AVT and high viral levels revealed higher AFP levels, larger tumor sizes, and higher rates of major hepatectomy, all of which have been confirmed as important influencing factors for PHLF. Further multivariate logistic regression analysis identified the high level of HBV DNA (cutoff value: 269 IU/mL) as a risk factor for PHLF in the entire patient cohort (P < 0.001) and the non-AVT cohort (P = 0.001). The lack of statistical significance for preoperative AVT in the multivariate analysis could be attributed to its high collinearity with the variable of HBV DNA level[37]. Further multivariate regression analysis without the variable of HBV DNA level showed that AVT is a protective factor against PHLF (P = 0.039).
Several prediction models for PHLF have been developed using diverse variables, such as demographic characteristics, surgical records, indocyanine green (ICG)-based measures, blood tests, and medical imaging[38]. Based on data extracted from our study cohort, we selected three commonly used scoring models as reference models for predicting PHLF. AUC is broadly known as a commonly used metric for assessing the performance of classification prediction models. However, the limitations of AUC have been underscored by previous studies[39-41]; for instance, improvements in sensitivity at a specific classification threshold may not necessarily translate into changes in AUC. As a consequence, solely computing ΔAUC may not fully capture the effect of a variable, including the HBV DNA levels, on the predictive capacity of scoring models. Risk reclassification indicators indicate the ability of predictive models to precisely categorize individuals based on clinical significance; these indicators have resolved the shortcomings of AUC, and their increase is vital in determining treatment decisions and priority treatment targets. Among them, NRI can quantify the correct movement in categories[42], and IDI is an indicator that integrates all the changes in the predicted value of risk, which can resolve the shortcomings of NRI and the defects of AUC[34]. Considering the absence of a clearly defined risk threshold for PHLF, the event rate of 0.15 was set as the threshold for risk stratification in our datasets[43].
To fully evaluate the importance of the HBV DNA level in predicting PHLF, three indicators were used to examine model performance, including ΔAUC, categorical NRI, and IDI. First, all models incorporating the HBV DNA level as a variable demonstrated improved AUC values, implying their improved predictive performance for PHLF (P < 0.05). Furthermore, the addition of the HBV DNA level to all three models significantly improved categorical NRI and IDI (P < 0.0001), indicating enhanced risk prediction capabilities for both event and non-event groups. As demonstrated by our findings, HBV DNA level with an optimal cutoff value of 269 IU/mL significantly improves the risk prediction performance of all reference scoring models for PHLF incidence. Eventually, we further confirmed the clinical practicability of the model using DCA. Generally, HBV DNA level can be considered a valuable and novel predictor of PHLF in patients with HBV-related HCC.
As popularly known, HBV reactivation often occurs in HBV-related HCC patients undergoing partial hepatectomy, which could further impair liver function impairment and even cause tumor recurrence, making it one of the main clinical limitations. While orally taking antiviral drugs including entecavir and tenofovir dipivoxil, HCC patients should regularly review the changes in liver function, AFP, and HBV DNA level after surgery. Since our specific criteria for PHLF identification were primarily based on relevant indicators on or after postoperative day 5, additional studies are essential to investigate the impact of AVT and HBV DNA levels on the long-term prognosis of HCC patients.
Limitations
This study has noteworthy limitations. First, it is a single-center cohort study. Secondly, it is retrospective in design with potential selection bias. Although diligent efforts were undertaken to address missing data through multiple imputations, patients without preoperative HBV DNA results were excluded from our analysis. Therefore, there is a need to include a more diverse and extensive patient cohort through a multi-center prospective study to validate our findings. Thirdly, the optimal critical value of HBV DNA for predicting the occurrence of PHLF was 269 IU/mL in this study cohort, which is not absolute. Finally, important variables for evaluating postoperative liver function or predicting PHLF were not analyzed, including ICG 15-minute retention rate and residual liver volume or proportion calculated based on medical imaging. As such, there is a need to holistically evaluate the predictive efficacy of the HBV DNA level for PHLF within a multimodal and multivariable predictive model.
CONCLUSION
In conclusion, patients with HBV-related HCC who received standard AVT before surgery have a lower incidence of PHLF than their counterparts without preoperative AVT. Moreover, our findings highlight a positive correlation between increased preoperative HBV DNA levels and an augmented risk of PHLF. The present study underscores the promising value of HBV DNA level as a biomarker for predicting PHLF, which can help in clinical decision-making.
Footnotes
Provenance and peer review: Unsolicited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Gastroenterology and hepatology
Country of origin: China
Peer-review report’s classification
Scientific Quality: Grade C
Novelty: Grade B
Creativity or Innovation: Grade B
Scientific Significance: Grade C
P-Reviewer: Murata K S-Editor: Li L L-Editor: A P-Editor: Zhao YQ
Johnson PJ, Berhane S, Kagebayashi C, Satomura S, Teng M, Reeves HL, O'Beirne J, Fox R, Skowronska A, Palmer D, Yeo W, Mo F, Lai P, Iñarrairaegui M, Chan SL, Sangro B, Miksad R, Tada T, Kumada T, Toyoda H. Assessment of liver function in patients with hepatocellular carcinoma: a new evidence-based approach-the ALBI grade.J Clin Oncol. 2015;33:550-558.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 1174][Cited by in F6Publishing: 1818][Article Influence: 181.8][Reference Citation Analysis (0)]
Huang G, Lai EC, Lau WY, Zhou WP, Shen F, Pan ZY, Fu SY, Wu MC. Posthepatectomy HBV reactivation in hepatitis B-related hepatocellular carcinoma influences postoperative survival in patients with preoperative low HBV-DNA levels.Ann Surg. 2013;257:490-505.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 76][Cited by in F6Publishing: 94][Article Influence: 8.5][Reference Citation Analysis (0)]
Rahbari NN, Garden OJ, Padbury R, Brooke-Smith M, Crawford M, Adam R, Koch M, Makuuchi M, Dematteo RP, Christophi C, Banting S, Usatoff V, Nagino M, Maddern G, Hugh TJ, Vauthey JN, Greig P, Rees M, Yokoyama Y, Fan ST, Nimura Y, Figueras J, Capussotti L, Büchler MW, Weitz J. Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS).Surgery. 2011;149:713-724.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 1224][Cited by in F6Publishing: 1613][Article Influence: 124.1][Reference Citation Analysis (0)]
Simpson AL, Adams LB, Allen PJ, D'Angelica MI, DeMatteo RP, Fong Y, Kingham TP, Leung U, Miga MI, Parada EP, Jarnagin WR, Do RK. Texture analysis of preoperative CT images for prediction of postoperative hepatic insufficiency: a preliminary study.J Am Coll Surg. 2015;220:339-346.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 38][Cited by in F6Publishing: 41][Article Influence: 4.1][Reference Citation Analysis (0)]
Li Z, Lei Z, Xia Y, Li J, Wang K, Zhang H, Wan X, Yang T, Zhou W, Wu M, Pawlik TM, Lau WY, Shen F. Association of Preoperative Antiviral Treatment With Incidences of Microvascular Invasion and Early Tumor Recurrence in Hepatitis B Virus-Related Hepatocellular Carcinoma.JAMA Surg. 2018;153:e182721.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 40][Cited by in F6Publishing: 80][Article Influence: 13.3][Reference Citation Analysis (0)]
