Predicting risk of post-hepatectomy liver failure in patients undergoing liver resection with controlled low central venous pressure

Abstract

BACKGROUND

Post-hepatectomy liver failure (PHLF), represents a serious complication after liver resection, significantly impacting the long-term outcomes for patients who undergo such surgeries. There exists a strong correlation between intraoperative hemorrhage and transfusion requirements with the development of PHLF. Presently, a combination of hepatic portal occlusion techniques alongside controlled low central venous pressure (CLCVP) methodologies is extensively employed to mitigate intraoperative bleeding. Nonetheless, limited studies have analyzed the risk factors for PHLF under CLCVP.

AIM

To develop and validate a nomogram that predicts the risk factors associated with the development of PHLF patients undergoing liver resection with CLCVP.

METHODS

We conducted a retrospective analysis of 285 patients who underwent hepatectomy for the first time and had no history of prior non-index abdominal surgeries, with hepatic inflow occlusion combined with CLCVP from January to December 2019 in Hunan Provincial People’s Hospital. Univariate and multivariate regression analyses were used to identify preoperative and intraoperative risk factors for PHLF. Eligible patients were randomly divided into training and validation groups in a 7:3 ratio, and a nomogram prediction model was constructed.

RESULTS

The incidence of PHLF in these patients was 22.46%. Multiple logistic analysis showed that preoperative serum albumin level, causes of liver resection (cancer or others), and cirrhosis were independent preoperative risk factors for PHLF (P < 0.05) and that only post-blocking blood potassium concentration was an independent intraoperative risk factor for PHLF (P < 0.05). Least absolute shrinkage and selection operator regression analysis revealed that preoperative serum albumin level, direct bilirubin level (DBIL), platelet count, causes of liver resection (cancer or others), and cirrhosis were significant predictors of PHLF. The nomogram risk prediction model based on preoperative serum albumin level, DBIL, platelet count, causes of liver resection (cancer or others), cirrhosis and post-blocking blood potassium concentration can better predict the occurrence of PHLF.

CONCLUSION

For patients undergoing liver resection with CLCVP, serum albumin level, DBIL, platelet count, causes of liver resection (cancer or others), and cirrhosis are independent preoperative risk factors for PHLF.

Key Words: Post-hepatectomy liver failure; Controlled low central venous pressure; Risk factors; Hepatectomy; Liver resection

Core Tip: This is the first study to establish and validate a model for predicting the risk of post-hepatectomy liver failure (PHLF) in patients undergoing hepatectomy with controlled low central venous pressure. In this work, we determined that serum albumin level, direct bilirubin level, platelet count, causes of liver resection (cancer or others), and cirrhosis were the main preoperative risk factors associated with PHLF in patients undergoing hepatectomy with controlled low central venous pressure. There is no evidence showing that intraoperative variables other than post-blocking blood potassium concentration may affect PHLF in these patients. The predictive model established in this study holds significant potential for enhancing the identification and risk stratification of patients, thereby aiding in postoperative management and improving clinical outcomes.

INTRODUCTION

Partial hepatectomy is the main treatment for a variety of liver and biliary diseases, but it also causes many complications, such as infection, biliary leakage, and gastrointestinal bleeding. The most important complication is post-hepatectomy liver failure (PHLF). It is a serious postoperative complication and is considered to be a marker of poor liver regeneration[1]. It is also one of the main causes of death after partial hepatectomy[2]. The incidence of PHLF ranges from 1.2% to 32%, and the mortality rate is approximately 1.6% to 2.8%[3-6].

Due to the liver’s rich blood supply, complex vascular distribution, numerous sinusoids, and high tissue fragility, it could become very difficult to control intraoperative bleeding, once the damage to hepatic vessels happens, especially the injuries of the hepatic and inferior vena cava. PHLF directly affecting patient outcomes, is related to both the amount of intraoperative bleeding and blood transfusions[7,8]. To reduce intraoperative bleeding, techniques such as hepatic portal occlusion combined with controlled low central venous pressure (CLCVP) technique were increasingly used in hepatectomy[8-10]. However, their effect on the incidence of PHLF is still controversial. Some studies point out that CLCVP can effectively reduce the amount of blood loss and blood transfusion during liver resection, and shorten the operation time[10,11]. There are many studies focused on the risk factors for PHLF, but fewer studies on the risk factors for PHLF under CLCVP. Therefore, based on the application of CLCVP in hepatectomy patients, we explored the preoperative and intraoperative risk factors for PHLF, and developed a nomogram in order to predict the incidence of PHLF. This model is good to figure out the possible effect of CLCVP on PHLF.

MATERIALS AND METHODS

Study population

A total of 1000 patients underwent liver resection for the first time from January to December 2019 in Hunan Provincial People’s Hospital. All patients underwent partial hepatectomy alone with intraoperative CLCVP [measured central venous pressure (CVP) ≤ 5 cm H₂O].

The inclusion criteria were: (1) Patients undergoing their first laparoscopic or open liver resection; (2) Age between 18 and 65 years; and (3) Preoperative assessment indicating suitability for surgery. The exclusion criteria included: (1) Emergency surgery; (2) Preoperative presence of significant organ dysfunction such as heart, kidney, or liver failure (Child-Pugh > C), untreated obstructive jaundice, or liver function not meeting surgical standards by the time of surgery; (3) History of major surgery or cancer; (4) Previous liver ablation or embolization; (5) History of radiotherapy or chemotherapy; (6) Combined non-liver surgery; and (7) Incomplete clinical data. According to the diagnostic criteria proposed by the International Liver Surgery Study Group in 2011, PHLF was defined as an increase in the international normalized ratio (INR) accompanied by hyperbilirubinemia on or after the fifth postoperative day[3].

The flowchart of the study is shown in Figure 1. Of the 1000 patients initially screened, 460 were excluded due to a history of prior abdominal surgeries, thus leaving 540 patients undergoing partial hepatectomy based on the CLCVP technique. Then 255 patients were excluded due to emergent surgery (n = 14), preoperative cardiac, renal, or liver failure (Child-Pugh > C) (n = 153), history of major surgery or other malignancies (n = 48), history of hepatic ablation or transcatheter arterial embolization (n = 23), radiotherapy or chemotherapy history (n = 11), non-biliary concomitant surgeries (n = 5), or incomplete clinical documentation (n = 1). Finally, 285 patients were enrolled. This study was conducted according to the Declaration of Helsinki and was approved by the Ethics Committee of Hunan Provincial People’s Hospital (The First Affiliated Hospital of Hunan Normal University) (IRB approval No. [2024]-273).

Figure 1 Flow chart of this study. CLCVP: Controlled low central venous pressure; PHLF: Post-hepatectomy liver failure; LASSO: Least absolute shrinkage and selection operator.

Variables

Preoperative risk factors: The following potential preoperative risk factors were evaluated: (1) General information: Age, height, weight, gender, history of hepatitis, liver cirrhosis, past medical history, etc.; (2) Lab Indicators: The last preoperative liver function, kidney function, heart function, complete blood count, blood lipids, myocardial enzymes, etc.; and (3) The results of re-examinations on or after the 5th postoperative day (if not checked on the 5^th day, the 4^th day’s results were taken; if the patient was discharged within 4 days postoperatively, the last re-examination results were taken);

Intraoperative risk factors: The following anesthesia and surgery-related variables were evaluated: Reason for liver resection, extent of resection, duration of surgery, extubation time, blood gas analysis before and after liver resection, blood loss, fluid intake, etc.

Outcome measure: PHLF was the outcome measure of this study, which was defined as total bilirubin (TBIL) and INR levels on or after the 5th postoperative day higher than preoperative levels or TBIL > 50 mmol/L (if the patient was discharged within 4 days postoperatively, liver function and coagulation function on the 4^th day were taken; if the patient was discharged within 4 days postoperatively and no liver function or coagulation function data was available on or after the 4^th day, the presence of postoperative liver failure was determined based on follow-up and outpatient reviews).

Statistical analysis

All statistical analyses were performed with SPSS software (version 25.0), the statistical software packages R (http://www.R-project.org, The R Foundation), and Free Statistics software versions 2.0. Normally distributed quantitative data are expressed as the mean ± SD, and comparisons between the two groups were performed using independent samples t-tests. Non-normally distributed quantitative data are expressed as median and interquartile range, and comparisons were performed using the Mann-Whitney U test. Qualitative data are displayed as n (%), and comparisons were made using the χ² test or Fisher’s exact test. Variables with P < 0.05 from (1); (2), and (3) were subjected to univariate logistic regression analysis, and variables significantly associated with the incidence of PHLF (P < 0.05) in the univariate logistic regression analysis were included in the multivariate regression analysis. The least absolute shrinkage and selection operator (LASSO) regression analysis was employed to identify predictor variables and construct a nomogram-based risk prediction model for PHLF. All statistical tests were two-sided, and a P-value of less than 0.05 was considered statistically significant.

RESULTS

Patient characteristics

There were 1000 patients who underwent their first hepatectomy with hepatic inflow occlusion combined with CLCVP at Hunan Provincial People’s Hospital between January and December 2019. Of these 1000 patients, 715 met the exclusion criteria and 285 were ultimately included in the analysis, as shown in Figure 1. The incidence of PHLF was 22.4% (64/285). A total of 141 (49.5) patients were male, and the median (interquartile range) age of patients was 54.0 (48.0-61.0) years (Table 1).

Table 1 Baseline characteristics in non-post-hepatectomy liver failure and post-hepatectomy liver failure groups, n (%).

Variable	Total (n = 285)	Non-PHLF (n = 221)	PHLF (n = 64)	P value
Preoperative risk factors
Gender	-	-	-	0.008
Male	141 (49.5)	100 (45.2)	41 (64.1)	-
Female	144 (50.5)	121 (54.8)	23 (35.9)	-
Age (years)	54.0 (48.0-61.0)	54.0 (47.0-61.0)	55.0 (49.0-62.0)	0.264
BMI (kg/m2), mean ± SD	22.5 ± 2.8	22.3 ± 2.8	22.9 ± 3.0	0.197
LOS (d), mean ± SD	18.1 ± 7.5	16.6 ± 5.9	23.2 ± 9.8	< 0.001
History of hepatitis	-	-	-	0.046
No	198 (69.5)	160 (72.4)	38 (59.4)	-
Yes	87 (30.5)	61 (27.6)	26 (40.6)	-
Cirrhosis	-	-	-	< 0.001
No	203 (71.2)	174 (78.7)	29 (45.3)	-
Yes	82 (28.8)	47 (21.3)	35 (54.7)	-
Jaundice	-	-	-	0.001
No	254 (89.1)	204 (92.3)	50 (78.1)	-
Yes	31 (10.9)	17 (7.7)	14 (21.9)
Ascites	-	-	-	0.263
No	252 (89.7)	197 (90.8)	55 (85.9)	-
Yes	29 (10.3)	20 (9.2)	9 (14.1)	-
Child-Pugh score	-	-	-	< 0.001
5	211 (74.0)	174 (78.7)	37 (57.8)	-
6	45 (15.8)	34 (15.4)	11 (17.2)	-
7	21 (7.4)	9 (4.1)	12 (18.8)	-
8	6 (2.1)	3 (1.4)	3 (4.7)	-
9	2 (0.7)	1 (0.5)	1 (1.6)	-
Cause of liver resection	-	-	-	< 0.001
Tumor	122 (42.8)	81 (36.7)	41 (64.1)	-
Other	163(57.2)	140 (63.3)	23(35.9)
TC (mmol/L)	4.4 ± 1.1	4.4 ± 1.0	4.2 ± 1.4	0.299
LDL-C (mmol/L)	2.6 (2.1-3.2)	2.6 (2.1-3.2)	2.4 (2.1-3.0)	0.206
TG (mmol/L)	1.2 (0.8-1.8)	1.2 (0.8-1.8)	1.0 (0.8-1.6)	0.515
Neutrophils (109/L)	3.4 (2.6-4.6)	3.5 (2.6-4.7)	3.2 (2.5- 4.4)	0.332
Lymphocytes (109/L), mean ± SD	1.6 ± 0.6	1.7 ± 0.6	1.4 ± 0.5	0.004
Leukocytes (109/L)	5.7 (4.6-7.1)	5.8 (4.7-7.2)	5.3 (4.3-6.7)	0.059
Platelets (109/L)	200.5 (149.0-264.0)	213.0 (163.0-264.0)	153.0 (108.0-249.5)	< 0.001
Hemoglobin (g/L)	128.4 ± 18.4	128.7 ± 16.4	127.3 ± 24.1	0.596
Albumin (g/L)	40.6 ± 5.1	41.2 ± 4.9	38.4 ± 5.1	< 0.001
TBIL (μmol/L)	14.9 (11.4-20.8)	14.5 (11.0-20.2)	17.0 (12.4-29.7)	0.012
DBIL (μmol/L)	5.0 (3.7-7.7)	4.8 (3.6-7.2)	5.5 (4.2-12.3)	0.005
ALT (U/L)	31.4 (19.0-56.4)	28.8 (18.0-50.7)	46.3 (28.1-66.5)	0.001
AST (U/L)	28.4 (21.8-45.2)	26.2 (20.8-39.2)	38.8 (28.7-52.6)	< 0.001
Intraoperative risk factors
Operation mode	-	-	-	0.611
Laparoscopic	153 (53.7)	122 (55.2)	31 (48.4)	-
Open	119 (41.8)	89 (40.3)	30 (46.9)	-
Conversion from laparoscopic to open	13 (4.6)	10 (4.5)	3 (4.7)	-
Extensive resection	-	-	-	0.09
No	160 (56.1)	130 (58.8)	30 (46.9)	-
Yes	125 (43.9)	91 (41.2)	34 (53.1)	-
Operation time (min)	327.7 ± 154.9	306.2 ± 134.4	401.8 ± 194.4	< 0.001
Liver blockage time (min)	41.0 (22.0-72.0)	38.0 (21.0-63.0)	58.0 (24.0-95.0)	0.003
Number of blocks	3.0 (1.0-4.0)	2.0 (1.0-4.0)	4.0 (2.0-6.0)	0.002
Total amount volume (mL), mean ± SD	2684.2 ± 1155.3	2490.7 ± 901.6	3352.6 ± 1610.2	< 0.001
Blood loss volume (mL)	200.0 (100.0-400.0)	200.0 (100.0-300.0)	300.0 (200.0-800.0)	< 0.001
Infusion of blood products	-	-	-	0.001
No	254 (89.1)	204 (92.3)	50 (78.1)	-
Yes	31 (10.9)	17 (7.7)	14 (21.9)	-
Urine volume (mL)	500.0 (300.0-950.0)	500.0 (300.0-800.0)	700.0 (362.5-1137.5)	0.001
Infusion of albumin	-	-	-	0.523
No	71 (24.9)	57 (25.8)	14 (21.9)	-
Yes	214 (75.1)	164 (74.2)	50 (78.1)	-
Pre-blocking PCO2 (mmHg)	42.0 (37.0-48.0)	43.0 (37.0-48.0)	42.0 (37.0-49.0)	0.91
Post-blocking PCO2 (mmHg)	43.0 (36.0-49.0)	43.5 (36.0-48.2)	43.0 (37.0-50.0)	0.535
Pre-blocking potassium (mmol/L)	3.7 (3.4-4.0)	3.6 (3.4-3.9)	3.7 (3.5-4.0)	0.15
Post-blocking potassium (mmol/L)	4.0 ± 0.6	3.9 ± 0.6	4.3 ± 0.6	< 0.001
Pre-blocking glucose (mmol/L)	5.8 (5.2-6.8)	5.8 (5.2-6.7)	6.2 (5.4-7.3)	0.25
Post-blocking glucose (mmol/L)	9.2 (7.0-11.1)	8.9 (7.0-11.0)	9.3 (7.0-11.0)	0.768
Pre-blocking lactic acid (mmol/L)	0.8 (0.6-1.2)	0.8 (0.6-1.2)	1.0 (0.7-1.4)	0.019
Post-blocking lactic acid (mmol/L)	1.6 (1.1-2.3)	1.4 (1.0-2.0)	2.0 (1.3-3.0)	< 0.001

PHLF: Post-hepatectomy liver failure; BMI: Body mass index; LOS: Length of stay; TC: Total cholesterol; LDL-C: Low-density lipoprotein cholesterol; TG: Triglycerides; TBIL: Total bilirubin; DBIL: Direct bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase.

Results of univariate and multivariate logistic regression analyses for risk factors for PHLF

We performed logistic regression analysis of the included variables. The univariate regression analysis revealed that among the potential preoperative risk factors, the following 12 variables were significantly associated with PHLF (P < 0.05): Preoperative serum albumin level (odds ratio [OR] = 0.89, 95 confidence interval [CI]: 0.84-0.95, P < 0.001), lymphocyte count (OR = 0.45, 95%CI: 0.26-0.78, P = 0.004), platelet count (OR = 0.99, 95%CI: 0.99-1, P = 0.001), prothrombin time (OR = 1.35, 95%CI: 1.09-1.67, P = 0.006), prothrombin activity level (OR = 0.99, 95%CI: 0.98-1, P = 0.043), TBIL (OR = 1.02, 95%CI: 1.01-1.03, P = 0.001), direct bilirubin (DBIL) (OR = 1.03, 95%CI: 1.01-1.05, P = 0.002), gender (OR = 2.16, 95%CI: 1.21-3.83, P = 0.009) history of viral hepatitis (OR = 1.79, 95%CI: 1.01-3.2, P = 0.048), jaundice (OR = 3.36, 95%CI: 1.55-7.27, P = 0.002), causes of liver resection (cancer or others) (OR = 0.32, 95%CI: 0.18-0.58, P < 0.001), and cirrhosis (OR = 4.47, 95%CI: 2.48-8.05, P < 0.001). Among the potential intraoperative risk factors, urinary volume (OR = 1.001, 95%CI: 1-1.001, P < 0.001), blood loss volume (OR = 1.002, 95%CI: 1.001-1.002, P < 0.001), infusion fluid volume (OR = 1.001, 95%CI: 1-1.001, P < 0.001), blood transfusion (OR = 3.36, 95%CI: 1.553-7.272, P = 0.002), operation time (OR = 1.004, 95%CI: 1.002-1.006, P < 0.001), hepatic portal occlusion time (OR = 1.014, 95%CI: 1.007-1.021, P < 0.001), number of blocks (OR = 1.255, 95%CI: 1.12-1.406, P < 0.001), post-blocking blood potassium concentration (OR = 2.222, 95%CI: 1.362-3.627, P = 0.001), and post-blocking blood lactate concentration (OR = 1.491, 95%CI: 1.164-1.908, P = 0.002) were significantly associated with PHLF (P < 0.05) (Table 2).

Table 2 Univariate and multivariate logistic analyses for preoperative risk factors.

Variable	Univariate logistic regression analysis		Multivariate logistic regression analysis
	OR, 95%CI	P value	OR, 95%CI	P value
Preoperative serum albumin level	0.89 (0.84-0.95)	< 0.001	0.93 (0.87-1)	0.036
Lymphocytes (109/L)	0.45 (0.26-0.78)	0.004	0.81 (0.41-1.61)	0.55
PLT (109/L)	0.99 (0.99-1)	0.001	1 (0.99-1)	0.212
Prothrombin time (s)	1.35 (1.09-1.67)	0.006	1.19 (0.8-1.77)	0.389
Prothrombin activity level (%)	0.99 (0.98-1)	0.043	1 (0.98-1.02)	0.822
TBIL (μmol/L)	1.02 (1.01-1.03)	0.001	1 (0.96-1.05)	0.917
DBIL (μmol/L)	1.03 (1.01-1.05)	0.002	1.01 (0.95-1.09)	0.697
Causes of liver resection, (cancer or others)	0.32 (0.18-0.58)	< 0.001	0.39 (0.16-0.94)	0.037
Gender	2.16 (1.21-3.83)	0.009	1.55 (0.75-3.19)	0.238
History of viral hepatitis	1.79 (1.01-3.2)	0.048	0.72 (0.29-1.82)	0.492
Cirrhosis	4.47 (2.48-8.05)	< 0.001	3.1 (1.56-6.17)	0.001
Jaundice	3.36 (1.55-7.27)	0.002	0.98 (0.28-3.42)	0.972

OR: Odds ratio; CI: Confidence interval; PLT: Platelets; TBIL: Total bilirubin; DBIL: Direct bilirubin.

The multiple logistic analysis showed that serum albumin level (OR = 0.93, 95%CI: 0.87-1, P = 0.036), causes of liver resection (cancer or others) (OR = 0.39, 95%CI: 0.16-0.94, P = 0.037), and cirrhosis (OR = 3.1, 95%CI: 1.56-6.17, P = 0.001) were independent preoperative risk factors for PHLF (P < 0.05) while only post-blocking blood potassium concentration (OR = 1.82, 95%CI: 1.05-3.17, P = 0.033) was an independent intraoperative risk factor for PHLF (P < 0.05) (Table 3).

Table 3 Univariate and multivariate logistic analyses for intraoperative risk factors.

Variable	Univariate logistic regression analysis		Multivariate logistic regression analysis
	OR, 95%CI	P value	OR, 95%CI	P value
Urinary production (mL)	1.001 (1-1.001)	< 0.001	1 (1-1)	0.593
Blood loss volume (mL)	1.002 (1.001-1.002)	< 0.001	1 (1-1)	0.567
Infusion fluid volume (mL)	1.001 (1-1.001)	< 0.001	1 (1-1)	0.148
Blood transfusion	3.36 (1.553-7.272)	0.002	0.94 (0.25-3.56)	0.93
Operation time (min)	1.004 (1.002-1.006)	< 0.001	1 (1-1)	0.954
Hepatic portal occlusion time (min)	1.014 (1.007-1.021)	< 0.001	1.01 (0.97-1.04)	0.641
Number of blocks	1.255 (1.12-1.406)	< 0.001	0.93 (0.55-1.57)	0.787
Post-blocking blood potassium (mmol/L)	2.222 (1.362-3.627)	0.001	1.82 (1.05-3.17)	0.033
Post-blocking blood lactate (mmol/L)	1.491 (1.164-1.908)	0.002	1.08 (0.8-1.46)	0.629

OR: Odds ratio; CI: Confidence interval.

Results of LASSO regression analysis of risk factors for PHLF

Eligible patients were randomly divided into training and validation groups in a 7:3 ratio. We used the LASSO algorithm to further refine a set of 12 preoperative variables from the univariate regression analysis (Table 4). Finally, they were reduced to five potential predictors: Preoperative serum albumin level, DBIL, platelet count, causes of liver resection (cancer or others), and cirrhosis. The LASSO coefficient trajectory path diagram and the cross-validation curve are showed in Figure 2 and Table 5.

Figure 2 Application of least absolute shrinkage and selection operator regression analysis for predictor variable selection. A: The least absolute shrinkage and selection operator coefficient trajectory path diagram, where each line represents the coefficient trajectory of an independent variable; B: The cross-validation curve for least absolute shrinkage and selection operator regression analysis, with the dashed line on the left indicating the log (λ) value corresponding to the minimum mean-square error. By selecting log (λ) at the minimum mean-square error, five variables with non-zero coefficients were identified.

Table 4 Baseline characteristics in training and validation groups, n (%).

Variable	Total (n = 285)	Training group (n = 199)	Validation group (n = 85)	P value
PHLF (%)	-	-	-	0.504
No	221	157 (78.9)	64 (75.3)	-
Yes	64	42 (21.1)	21 (24.7)	-
Preoperative risk factors
Gender	-	-	-	0.98
Male	141 (49.5)	101 (50.8)	43 (50.6)	-
Female	144 (50.5)	98 (49.2)	42 (49.4)	-
Age (years)	54.0 (48.0-61.0)	54.0 (48.0-61.0)	55.0 (49.0-62.0)	0.262
BMI (kg/m2), mean ± SD	22.5 ± 2.8	22.4 ± 2.8	22.6 ± 2.9	0.551
LOS (d), mean ± SD	18.1 ± 7.5	17.7 ± 7.0	19.0 ± 8.5	0.175
History of hepatitis	-	-	-	0.266
No	198 (69.5)	142 (71.4)	55 (64.7)	-
Yes	87 (30.5)	57 (28.6)	30 (35.3)	-
Cirrhosis	-	-	-	0.52
No	203 (71.2)	140 (70.4)	63 (74.1)	-
Yes	82 (28.8)	59 (29.6)	22 (25.9)	-
Jaundice	-	-	-	0.474
No	254 (89.1)	179 (89.9)	74 (87.1)	-
Yes	31 (10.9)	20 (10.1)	11 (12.9)	-
Ascites	-	-	-	0.572
No	252 (89.7)	180 (90.5)	75 (88.2)	-
Yes	29 (10.3)	19 (9.5)	10 (11.8)	-
Child-Pugh score		-	-	0.198
5	211 (74.0)	144 (72.4)	66 (77.6)	-
6	45 (15.8)	35 (17.6)	10 (11.8)	-
7	21 (7.4)	16 (8)	5 (5.9)	-
8	6 (2.1)	4 (2)	2 (2.4)	-
9	2 (0.7)	0 (0)	2 (2.4)	-
Cause of liver resection	-	-	-	0.041
Tumor	122 (42.8)	77 (38.7)	44 (51.8)	-
Other	163(57.2)	122 (61.3)	41 (48.2)	-
TC (mmol/L), mean ± SD	4.4 ± 1.1	4.3 ± 1.1	4.4 ± 1.1	0.625
LDL-C (mmol/L)	2.6 (2.1-3.2)	2.5 (2.1-3.2)	2.7 (2.1-3.0)	0.816
TG (mmol/L)	1.2 (0.8-1.8)	1.2 (0.8-1.7)	1.1 (0.8-1.9)	0.941
Neutrophils (109/L)	3.4 (2.6-4.6)	3.5 (2.6-4.6)	3.4 (2.6-4.6)	0.698
Lymphocytes (109/L)	1.6 ± 0.6	1.6 ± 0.6	1.6 ± 0.5	0.447
Leukocytes (109/L)	5.7 (4.6-7.1)	5.7 (4.7-7.2)	5.6 (4.5-7.0)	0.738
Platelets (109/L)	200.5 (149.0-264.0)	199.0 (149.0-259.0)	206.0 (156.0-279.0)	0.587
Hemoglobin (g/L)	128.4 ± 18.4	127.8 ± 18.1	129.9 ± 19.1	0.382
Albumin (g/L)	40.6 ± 5.1	40.4 ± 5.2	41.0 ± 4.8	0.394
TBIL (μmol/L)	14.9 (11.4-20.8)	14.8 (10.9-22.4)	15.2 (11.9-19.8)	0.873
DBIL (μmol/L)	5.0 (3.7-7.7)	5.1 (3.7-7.8)	4.8 (4.0-7.6)	0.969
ALT (U/L)	31.4 (19.0-56.4)	31.5 (19.6-59.0)	30.0 (18.5-46.3)	0.230
AST (U/L)	28.4 (21.8-45.2)	28.9 (21.6-46.0)	28.1 (22.0-44.6)	0.905
Intraoperative risk factors
Operation mode	-	-	-	0.378
Laparoscopic	153 (53.7)	107 (53.8)	46 (54.1)	-
Open	119 (41.8)	85 (42.7)	33 (38.8)	-
Conversion from laparoscopic to open	13 (4.6)	7 (3.5)	6 (7.1)	-
Extensive resection	-	-	-	0.416
No	160 (56.1)	109 (54.8)	51 (60)	-
Yes	125 (43.9)	90 (45.2)	34 (40)	-
Operation length (min), mean ± SD	327.7 ± 154.9	322.1 ± 155.8	340.5 ± 153.7	0.360
Liver blockage time (min)	41.0 (22.0-72.0)	40.0 (18.2-68.0)	43.0 (28.0-75.0)	0.294
Number of blocks	3.0 (1.0-4.0)	3.0 (1.0-4.0)	3.0 (2.0-5.0)	0.520
Total amount volume (mL), mean ± SD	2684.2 ± 1155.3	2696.7 ± 1134.0	2647.2 ± 1214.2	0.742
Blood loss volume (mL)	200.0 (100.0-400.0)	200.0 (100.0-400.0)	200.0 (150.0-400.0)	0.102
Infusion of blood products	-	-	-	0.344
No	254 (89.1)	175 (87.9)	78 (91.8)	-
Yes	31 (10.9)	24 (12.1)	7 (8.2)	-
Urine volume (mL)	500.0 (300.0-950.0)	500.0 (300.0-912.5)	500.0 (300.0-1000.0)	0.995
Infusion of albumin	-	-	-	0.752
No	71 (24.9)	48 (24.1)	22 (25.9)	-
Yes	214 (75.1)	151 (75.9)	63 (74.1)	-
Pre-blocking PCO2 (mmHg)	42.0 (37.0-48.0)	42.0 (37.0-49.0)	42.0 (35.0-47.0)	0.480
Post-blocking PCO2 (mmHg)	43.0 (36.0-49.0)	43.0 (36.0-49.0)	44.0 (36.0-49.5)	0.805
Pre-blocking potassium (mmol/L)	3.7 (3.4-4.0)	3.6 (3.3-3.9)	3.7 (3.5-4.0)	0.072
Post-blocking potassium (mmol/L), mean ± SD	4.0 ± 0.6	4.0 ± 0.6	4.1 ± 0.7	0.082
Pre-blocking glucose (mmol/L)	5.8 (5.2-6.8)	5.8 (5.2-6.8)	5.8 (5.2*6.8)	0.876
Post-blocking glucose (mmol/L)	9.2 (7.0-11.1)	8.9 (7.0-10.8)	9.5 (7.5*11.4)	0.181
Pre-blocking lactic acid (mmol/L)	0.8 (0.6-1.2)	0.8 (0.6-1.2)	0.8 (0.7-1.2)	0.744
Post-blocking lactic acid (mmol/L)	1.6 (1.1-2.3)	1.5 (1.0-2.2)	1.6 (1.2-2.5)	0.284

PHLF: Post-hepatectomy liver failure; BMI: Body mass index; LOS: Length of stay; TC: Total cholesterol; LDL-C: Low-density lipoprotein cholesterol; TG: Triglycerides; TBIL: Total bilirubin; DBIL: Direct bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase.

Table 5 Variable filtering.

Number	Variable, name	Coefficient	Lambda, type	Lambda, P value
0	Intercept	-0.63	Lambda1se	0.049
1	Gender	-	Lambda1se	0.049
2	History of viral hepatitis	-	Lambda1se	0.049
3	Cirrhosis	0.768	Lambda1se	0.049
4	Jaundice	-	Lambda1se	0.049
5	Causes of liver resection (cancer or others)	-0.37	Lambda1se	0.049
6	Lymphocyte count (109/L)	-	Lambda1se	0.049
7	PLT (109/L)	0	Lambda1se	0.049
8	Preoperative serum albumin level	-0.028	Lambda1se	0.049
9	TBIL (μmol/L)	-	Lambda1se	0.049
10	DBIL (μmol/L)	0.01	Lambda1se	0.049
11	Prothrombin time (second)	-	Lambda1se	0.049
12	Prothrombin activity level (%)	-	Lambda1se	0.049

PLT: Platelets; BIL: Total bilirubin; DBIL: Direct bilirubin.

Development and validation of a nomogram

We selected the five preoperative variables including preoperative serum albumin level, DBIL, platelet count, causes of liver resection (cancer or others), and cirrhosis based on LASSO regression analysis to develop a nomogram (Figure 3). Using the nomogram, the scoring is as follows: 47.5 points for preoperative serum albumin level of 5 g/L, 50 points for platelet count of 350 × 10⁹/L, 21 points for cirrhosis, 11 points for causes of liver resection (cancer or others), and 15 points for DBIL level of 40 μmol/L, resulting in a total score of 200, which corresponds to an event risk value of 0.9.

Figure 3 Nomogram constructed for predicting post-hepatectomy liver failure. PLT: Platelets; DBIL: Direct bilirubin.

We performed a detailed sensitivity and specificity analysis of the model, with the findings visually represented in box plots (Figure 4A). Furthermore, the calibration curve demonstrated a strong alignment between the predicted and actual values for PHLF, as illustrated in Figure 4B. In addition, the decision curve analysis validated the clinical utility of the nomogram, revealing a favorable net benefit across various threshold probabilities, which underscores its practical applicability in clinical settings (Figure 4C).

Figure 4 Validation of the nomogram: A: Validation using the area under the receiver operating characteristics curve; B: Calibration curve for verifying the consistency between the predicted probabilities and the actual observed probabilities; C: Decision curve analysis to assess the clinical efficacy of the model. AUC: Area under the receiver operating characteristics curve.

DISCUSSION

The incidence of liver cancer and severe intrahepatic bile duct stones is relatively high in China. The patients with liver cancer in China account for about 45% of the world’s cases, and the incidence of intrahepatic bile duct stones accounts for about 16.1%[12], most of which require liver resection. The main postoperative risk of hepatectomy is PHLF. According to previous reports in the literature, PHLF is closely related to prognosis and is the most important cause of perioperative mortality[13]. PHLF has not yet had a uniform definition. Balzan et al[14] proposed the “50-50” standard - the prothrombin index < 50% and serum total bilirubin > 50 μmol/L. However, this standard has a high specificity but low sensitivity and did not consider other clinical indicators, limiting its clinical application. In 2011, the International Liver Surgery Research Group proposed a new definition and severity grading criteria for PHLF, which is defined as impairment of postoperative liver synthesis, secretion, and detoxification functions, mainly using total bilirubin and INR as the evaluation criteria[3]. Using the International Liver Surgery Research Group criteria, in this study, out of 285 patients undergoing liver resection, 64 developed PHLF, with an incidence of 22.4%, which is generally consistent with the incidence reported in the literature[4,5].

Hepatoportal block is an important method to reduce intraoperative bleeding and improve the success rate of liver resection, but excessive liver blockage time can cause aggravated ischemia-reperfusion injury and increase the risk of PHLF[15,16]. Due to the abundant blood supply in the liver, the portal block alone cannot prevent the bleeding of the hepatic vein, and the filled intrahepatic vein will also increase the difficulty of resecting liver parenchyma. Therefore, the combined application of CLCVP technique to maintain low CVP and collapse the vena cava and its branches, can effectively reduce wound bleeding, improve the surgical field, conducive to the free of liver parenchyma, and indirectly reduce the total blocking time. CLCVP is an anesthetic technique that uses anesthesia-related drugs and techniques to maintain the CVP at 5 cm H₂O and below, while ensuring the systolic blood pressure above 90 mmHg or the average arterial pressure above 60 mmHg to reduce intraoperative bleeding. However, the effect of intraoperative application of CLCVP on the occurrence of PHLF remains unclear. Wang et al[17] found that CLCVP is not only easy to achieve in anesthesia management, but also helps to reduce blood loss during liver resection, shorten hospital stay, and have no obvious effect on liver and kidney function. A study of the best CVP in liver resection in 97 hepatocellular carcinoma patients found that the intraoperative blood loss also decreased simultaneously as CVP decreased[18]. However, some studies have pointed out that excessively low CVP maintenance level should be avoided to decrease the risk of tissue malperfusion and PHLF, especially in elderly or prolonged portal blockade patients[19]. In this study, we found that for patients treated with the CLCVP technique, neither the liver clamping time nor the blood loss was a risk factor for PHLF. This may be related to the fact that the application of CLCVP reduces bleeding and accelerates the surgical process. However, the serum potassium level after clamping was a significant risk factor for PHLF (P < 0.05). This indicates that when applying CLCVP, attention must be paid to electrolyte changes, especially the change in serum potassium, and intraoperative intervention should be carried out promptly to reduce the incidence of PHLF.

In addition to intraoperative factors, the occurrence of PHLF is also closely related to patients’ preoperative basic factors and liver conditions. Previous studies have shown that the preoperative risk factors for PHLF cover patients’ own factors (such as age, gender, and body mass index)[20], as well as factors related to impaired liver reserve function (such as cholestasis, decreased secretory function, liver cirrhosis, and platelet abnormalities)[3,21,22]. In China, the incidence of liver cancer and hepatobiliary stones is relatively high. Most patients have different degrees and types of liver cirrhosis when undergoing surgery. A large number of studies have confirmed that hypoalbuminemia can have a serious impact on the success rate and prognosis of hepatectomy[23] and a retrospective study involving 268 patients who underwent hepatectomy found that the preoperative albumin level is an important indicator for predicting liver failure after hepatectomy.

In this study, through multivariate logistic regression analysis, it was found that based on the CLCVP technique, liver cirrhosis is a significant risk factor for PHLF in patients undergoing hepatectomy (P = 0.001). Therefore, for patients with liver cirrhosis, performing hepatectomy requires more thorough preoperative preparation, more refined intraoperative anesthesia management, and more meticulous postoperative management. In our study, it was found that preoperative albumin is also an independent risk factor for PHLF in patients undergoing hepatectomy with the application of the CLCVP technique (P = 0.036), and it has a negative linear correlation with the occurrence of PHLF. Therefore, before performing hepatectomy, close attention should be paid to the patient’s albumin level. For patients with hypoproteinemia, reasonable supplementation of human albumin during the perioperative period may have a positive effect in improving the success rate of the surgery and reducing the incidence of postoperative complications. Moreover, in our research, patients with liver cancer as the indication for surgery emerged as a notable risk factor for PHLF (P = 0.037). When compared with patients undergoing surgery for other causes, those with liver cancer exhibit a markedly more rapid decline in liver function, possess poor hepatic reserve capacity, and necessitate a more extensive scope of liver resection[24].

CONCLUSION

In a retrospective analysis of 285 patients undergoing hepatectomy with CLCVP, cirrhosis, preoperative albumin level, causes of liver resection (cancer or others), DBIL and platelet count were independent risk factors and predictors of PHLF in hepatectomy patients under CLCVP, and this model has a good ability to predict PHLF in hepatectomy patients under CLCVP. Cirrhosis, liver cancer, low preoperative albumin, and high preoperative DBIL can lead to an increased likelihood of PHLF. This suggests that the risk of PHLF is higher in patients with cirrhosis using the CLCVP technique. Before surgery, reasonably increasing albumin levels and reducing DBIL levels may have a positive impact in reducing the risk of PHLF. In this study, we found that for patients treated with the CLCVP technique, neither the liver clamping time nor the blood loss was a risk factor for PHLF. This may be related to the fact that the application of CLCVP reduces bleeding and accelerates the surgical process.