Risk Stratification Based on Acute-on-chronic liver failure in cirrhotic patients hospitalized with acute variceal bleeding

doi:10.21203/rs.3.rs-2090435/v1

Background and aims: Acute variceal bleeding (AVB) is a life-threatening complication of cirrhosis. This study aimed to evaluate the role of Acute-on-chronic liver failure (ACLF) in the risk stratification of cirrhotic patients hospitalized with AVB.

Methods: Prospective data of 417 consecutive cirrhotic patients hospitalized with AVB were retrospectively collected and analyzed. ACLF was defined by European Association for the Study of Liver-Chronic Liver Failure Consortium (EASL-CLIF-C) and diagnosed/graded with chronic liver failure-organ failure score.

Results: A total of 218 (52.2%) patients were diagnosed with EASL-ACLF (grade 1: 18.3%, grade 2: 32.2%, grade 3: 49.5%) at admission. The cumulative 6-week survival rate in patients with ACLF was significantly lower than those without (53.2% vs. 88.9%, P<0.001) and decreased in line with the severity of ACLF (P <0.001). In multivariate analysis, presence of ACLF remained as an independent risk factor for 6-week mortality after adjusting for confounding factors (HR =2.74, p<0.001). CLIF-C ACLF and CLIF-C AD outperformed traditional prognosis scores in the prediction of 6-week mortality in patients with and without ACLF, respectively.

Conclusion: ACLF at admission is an independent predictor for the 6-week mortality in cirrhotic patients with AVB. For AVB patients with and without ACLF, CLIF-C ACLF and CLIF-C AD outperformed other prognostic scores in the prediction of 6-week mortality, respectively.

Acute Variceal Bleeding

Acute-on-chronic liver failure

Risk stratification

Prognosis

Acute variceal bleeding (AVB) is one of the most common and serious complications of cirrhosis. Despite recent improvement in therapy (medication, endoscopy), up to 10–15% of patients still have persistent bleeding or early rebleeding and the overall mortality with each episode of AVB is still approximately 15% to 25% at six weeks [^1][2][3]. Early identification of these high-risk patients and taking alternative more effective treatments, such as preemptive transjugular intrahepatic portosystemic shunt (p-TIPS), were recommended by multiple international consensus to improve the prognosis of these patients [^4][5][6]. So far, factors found to be associated with the poor prognosis in cirrhotic patients with AVB include shock on admission, hepatic venous pressure gradient (HVPG) >20 mmHg, concurrence of hepatic encephalopathy (HE) or hepatocellular carcinoma (HCC), renal failure and bacterial infection, etc [^{7][8][9][10][11]}. Some prognosis scores, such as Child-Turcotte-Pugh (CTP) [^12], Model for End-stage Liver Disease (MELD) [^13]and MELD-Na [^14] have been proposed for predicting the prognosis of these patients. However, as these scores mainly reflect the severity of liver disease, their prediction accuracy may decrease when they encounter extrahepatic organ failures which have a vital impact on the prognosis of these patients [^15][16].

Acute-on-chronic liver failure (ACLF) was first defined by EASL (European Association for the Study of the Liver) as a syndrome characterized by acute decompensation (ie, ascites, encephalopathy, AVB, bacterial infection) of cirrhosis, multisystem organ failures and high short-term mortality [^17]. In this study, the 28-day mortality in cirrhotic patients with ACLF is significantly higher than that in patients without ACLF (33.9% VS. 4.7%, P <0.001). Nowadays, with the improvement in hemostatic technology, the rate of cirrhotic patients with AVB who died of hemorrhagic shock gradually decreased. Instead, most patients died of liver failure or multiple organ failures ^[17]. Although AVB is a well recognized precipitant leading to the occurrence and development of ACLF [^18][19][20], the role of ACLF in the prognosis of cirrhotic patients with AVB has not yet been fully investigated.

In this study, we aimed at addressing the following 3 clinically relevant issues: (1) whether the presence and grade of ACLF at admission was independently associated with the poor prognosis in cirrhotic patients hospitalized with AVB. (2) analyse the difference of demographic characteristics, clinical features and laboratory parameters between patients with and without ACLF. (3) identify the reliable prognosis scores in ACLF and mere AD (acute decompensation, without ACLF) patients, respectively.

Patients

The medical record of patients enrolled in this retrospective cohort study were obtained from the Medical Information Mart for Intensive Care (MIMIC-Ⅳ ) database (version 2.0) [^21][22], which is a large, freely-available database comprising deidentified health-related data from patients admitted to the critical care units of the Beth Israel Deaconess Medical Center between 2008 and 2019. Inclusion criteria: cirrhotic patients hospitalized with acute variceal bleeding. Exclusion criteria: (1) Age <18 years old. (2) without ICU stays. (3) Incomplete records. The access to the MIMIC-IV database was approved by the Institutional Review Board of the Beth Israel Deaconess Medical Center and Massachusetts Institute of Technology after the completion of online course and examination. Given the public availability of MIMIC-Ⅳ database, with private information of all patients being anonymized, the local ethics committee’s approval was waived.

Data collection

We obtained the medical record information of each enrollment through the MIMIC-Ⅳ database (version 2.0), which was released on June 12, 2022 including in and out-of-hospital date of death. The medical data of patients in MIMIC-Ⅳ 2.0 was extracted through Postgres Structured Query Language (PostgreSQL) programming in Navicat Premium (version 15.0.12). We first retrieved the subject_id and hadm_id (hospital admission_id) of patients with acute variceal bleeding and cirrhosis by searching the International Classification of Disease (ICD)_code of (4560, 45620) and (5712, 5715, 5716), respectively. After that, the subject_id and hadm_id of cirrhotic patients with acute variceal bleeding was retrieved by taking the intersection of these two cohorts. And then, we intersected the subject_id and hadm_id of each cirrhotic patients with AVB and those of all ICU stayed patients to get the ICU stay_id of cirrhotic patients with AVB. Finally, we extracted the medical information of each enrollment from the corresponding tables, including age, gender, race, comorbidities, first day vital signs, first day laboratory parameters, vasopressors therapy, mechanical ventilation therapy, renal replacement treatment (RRT) therapy, severity scores, and survival information, etc. For those admitted multiple times to the ICU, the data for their first admissions were used. For the laboratory tests measured more than one time, we selected the maximum or minimum of them according to their clinical implications. For example, the white blood cell and total bilirubin were selected with the maximum of them, while the albumin and serum sodium were selected with the minimum of them. For all the enrolled patients, we calculated the common prognosis scores (CTP, MELD and MELD-Na). As CLIF-C ACLF and CLIF-C AD were the specific prognosis scores for ACLF and mere AD patients, respectively, we additionally calculated them. The MELD score can be directly obtained from MIMIC Ⅳ database. The MELD-Na score was calculated using the following formula: MELD-Na= MELD + 1.59 (135-Na) with maximum and minimum Na of 135 and 120 mEq/L, respectively ^[14]. The CLIF-C ACLF and CLIF-C AD score were calculated using the following formulas: CLIF-C ACLF = 10 × [0.33 × CLIF-OFs + 0.04 × Age + 0.63 × ln(WBC count) − 2] [^23], CLIF-C ADs = 10×[0.03×Age(year)+0.66×Ln(Creatinine(mg/dL)+1.71×Ln(INR)+ 0.88×Ln(WBC{10⁹cells/L}) -0.05×Sodium(mmol/L) +8] [^24]. CTP ^[12]and CLIF-OF^[23]were calculated as described previously. ACLF was defined by European Association for the Study of Liver-Chronic Liver Failure Consortium (EASL-CLIF) ^[17] and diagnosed / graded according to chronic liver failure-organ failure (CLIF-OF) score. Specifically, liver failure was defined by total bilirubin ≥12mg/dl; coagulation failure was defined by INR ≥2.5; kidney failure was defined by creatinine >2mg/dL or renal replacement therapy; circulatory failure was considered when vasopressor therapy was needed to maintain blood pressure; respiratory failure was diagnosed when PaO2 /FiO2 ≤200 or SpO2 /FiO2 ≤214 or mechanical ventilation was required for reasons other than airway protection and in the absence of HE grade III or IV; Brain failure was defined by HE grade III or IV (West Haven). Grade I ACLF was defined as: (1) having single kidney failure; (2) single failure of the liver, coagulation, circulation or respiration along with a serum creatinine level ranging from 1.5 to 1.9 mg/dL and/or mild to moderate hepatic encephalopathy; (3) single cerebral failure along with a serum creatinine level ranging from 1.5 and 1.9 mg/dL; Grade II as having two organ failures; Grade III as ≥3 organ failures.

Study outcomes

The outcome of this study was 6-week all-cause mortality according to the Baveno VI Consensus Workshop[3] .The start date of follow-up was the date of the patient’s admission. The follow-up started on the date of patient’s admission and ended at 6 weeks later or the date of patient’s death.

Statistical analysis

Continuous variables with normal or skew distribution were described as means (± standard deviation) and medians (interquartile range) , respectively. Categorical variables were described as numbers (percentage). Univariate and multivariate Cox-proportional hazards regression analysis were performed to identify the risk factors for the 6-week morality of enrolled patients. Variables with p values <0.05 in the univariate analysis were considered for the multivariate analysis with backward stepwise method. To evaluate the performance of prognostic scores in predicting the 6-week mortality, discrimination, calibration and overall performance of each score were studied. Discrimination refers to the ability of stratifying patients according to their risk of developing the outcome, whereas calibration refers to the ability of predicting absolute risks (how closely the predicted probabilities agree with the actual outcomes). The discrimination performance of prognostic scores were evaluated by performing receiver operating characteristics (ROC) curve and calculating the area under ROC (AUROC). Discrimination performance of the prognostic scores were compared by Delong test. The calibration performance was evaluated by performing Hosmer-Lemeshow goodness-of-fit test and plotting the calibration curves to visually inspect the observed versus predicted probabilities of death. The overall performance was assessed by testing the Brier score and R²value. A lower Brier score or higher R²value indicate a better overall performance. Cumulative survival curves were plotted using the Kaplan–Meier method and compared using the log-rank test. Statistical analysis was performed using SPSS software version 22.0 (IBM Corp, Armonk, NY, USA) , Medcalc software version 19.0.4 (MedCalc Software, Belgium), STATA 15.0 (Statistics/Data Analysis, Stata Corp, USA) and R version 4.2.0 (The R Foundation for Statistical Computing). A two-tailed P value <0 .05 was considered to be statistically significant.

Baseline characteristics

A total of 603 consecutive patients with acute variceal bleeding (AVB) were screened, and 186 patients were excluded for the following reasons: without cirrhosis (n =46), without ICU stays (n =129) and incomplete records (n =11). Finally, 417 patients with cirrhosis and AVB who met the inclusion and exclusion criteria were included in this study. Their baseline characteristics were shown in Table 1. Patients were predominantly male 297 (71.2%), with a median age of 57 years. Patients were predominantly white 277 (66.4%). The main etiology of cirrhosis was alcohol (56.8%). With regard to the complications of cirrhosis, a total of 244 (58.5%), 247 (59.3%), 109 (26.1%) and 49 (11.7%) patients had ascites, hepatic encephalopathy (HE), bacterial infection and portal vein thrombosis (PVT) at admission, respectively. 52 (12.4%) patients had concurrent hepatocellular carcinoma (HCC). Of all the patients included, 218 (52.2%) had EASL-ACLF at baseline (40 [18.3%], 70 [32.2%] and 108 [49.5%] had ACLF grade 1, grade 2 and grade 3, respectively). Age, gender, race and etiology were similar between patients with and without ACLF. As expected, patients with ACLF more frequently presented with ascites, bacterial infections and hepatic encephalopathy, as well as significantly higher heart rate, peripheral white blood cell (WBC) count, serum total bilirubin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), international normalized ratio (INR), prothrombin time (PT), serum creatinine, blood urea nitrogen, serum potassium, glucose and lower mean arterial pressure, SPO2 (percutaneous oxygen saturation)/ FIO2 (fraction of inspired oxygen), hemoglobin, albumin and serum sodium at baseline. In addition, the ACLF patients significantly presented higher prognosis scores (CTP, MELD, MELD-Na) and proportion of CTP class C than mere AD patients (P <0.001). There were no significant differences of the incidence of comorbidities between ACLF and mere AD patients, except renal disease (P =0.005). Besides, the length of ICU stay and hospital stay in ACLF patients were significantly higher than that in mere AD patients.

Table 1 Baseline characteristics of all patients and patients with or without ACLF (N = 417)

Baseline characteristics	All patients (n= 417)	No ACLF (n= 199)	ACLF (n= 218)	P value
Gender n (%)				0.305
Male	297 (71.2)	137 (68.8)	160 (73.4)
Female	120 (28.8)	62 (31.2)	58 (26.6)
Age	57 (49-63)	58 (50-63)	56 (49-63)	0.406
Race n (%)				0.209
WHITE	277 (66.4)	140 (70.4)	137 (62.8)
BLACK	48 (11.5)	16(8.0)	32 (14.7)
HISPANIC/LATINO	45 (10.8)	23 (11.6)	22 (10.1)
ASIAN	19 (4.6)	9 (4.5)	10 (4.6)
OTHERS	28 (6.7)	11 (5.5)	17 (7.8)
Etiology of cirrhosis n (%)				0.279
Alcohol	237 (56.8)	104 (52.3)	133 (61.0)
Virus	29 (7.0)	17 (8.5)	12 (5.5)
Alcohol+Virus	18 (4.3)	10 (5.0)	8 (3.7)
Others	133 (31.9)	68 (34.2)	65 (29.8)
Concurrent HCC n (%)	52 (12.5)	31 (15.6)	21 (9.6)	0.066
Comorbidities
Miocardial infarction	18 (4.3)	6 (3.0)	12 (5.5)	0.212
Congestive heart failure	32 (7.7)	14 (7.0)	18 (8.3)	0.640
Cerebrovascular disease	15 (3.6)	6 (3.0)	9 (4.1)	0.542
Chronic pulmonary disease	64 (15.3)	31 (15.6)	33 (15.1)	0.901
Diabetes	29 (7.0)	13 (6.5)	16 (7.3)	0.746
Peripheral vascular disease	13 (3.1)	7 (3.5)	6 (2.8)	0.653
Renal disease	48 (11.5)	14 (7.0)	34 (15.6)	0.006
Decompensation at admission
Ascites n (%)	244 (58.5)	82 (41.2)	162 (74.3)	<0.001
HE n (%)	247 (59.3)	77 (38.7)	170 (78.0)	<0.001
Ⅰ+Ⅱ	131 (31.5)	56 (28.1)	75 (34.4)
Ⅲ+Ⅳ	116 (27.8)	21 (10.6)	95 (43.6)
Bacterial infection	109 (26.1)	29 (14.6)	80 (36.7)	<0.001
Pneumonia	39 (9.3)	13 (6.5)	26 (11.9)
Urinary tract infection	38 (9.1)	10 (5.0)	28 (12.8)
Spontaneous bacterial peritonitis	32 (7.6)	7 (3.5)	25 (11.5)
Others	16 (3.8)	5 (2.5)	11 (5.0)
PVT	49 (11.8)	26 (13.1)	23 (10.6)	0.426
Vital signs
Mean arterial pressure (mmHg)	75 (69-82)	79 (71-86)	72 (67-78)	<0.001
Heart rate (bpm)	86 (75-98)	85 (73-98)	87 (76-98)	0.005
SPO2/FIO2	405 (150-443)	433 (192-448)	229 (104-438)	<0.001
Laboratory tests
White blood cell (10⁹/L)	10.6 (6.8-16.1)	7.7 (5.6-11.3)	13.9 (9.6-20.0)	<0.001
Hemoglobin (mg/dl)	8.3±1.9	8.6±1.8	8.0±1.9	0.001
Platelet (10⁹/L)	71.0 (48.0-108.0)	72.0 (50.5-113.0)	68.5 (45.0-106.8)	0.079
Total bilirubin (mg/dL)	3.2 (1.6-7.3)	2.0 (1.2-4.0)	5.2 (2.4-13.2)	<0.001
Albumin, g/dL	2.9±0.6	3.0 (2.6-3.4)	2.8±0.7	<0.001
ALT (U/L)	39 (24-65)	35 (24-52)	42 (24-100)	0.005
AST (U/L)	76 (46-155)	64 (43-112)	93 (53-262)	<0.001
INR	1.8 (1.5-2.2)	1.6 (1.4-1.8)	2.1 (1.7-2.7)	<0.001
Prothrombin time (s)	19.2 (16.0-23.8)	17.1 (15.2-19.7)	22.7 (18.4-29.6)	<0.001
Serum creatinine, mg/dL	1.1 (0.8-1.9)	0.8 (0.7-1.0)	1.8 (1.1-2.8)	<0.001
Blood urea nitrogen (mg/dl)	29.0 (18.0-45.0)	22.0 (15.0-34.0)	37.0 (23.0-57.0)	<0.001
Serum sodium, mEq/L	137.0 (133.0-140.0)	138.0 (135.0-141.0)	136.0 (131.0-140.0)	0.001
Serum potassium, mEq/L	4.5 (4.1-5.4)	4.4 (4.0-4.9)	4.8 (4.2-5.7)	<0.001
Glucose (mg/dl)	149.0 (121.0-198.5)	137.0 (118.0-187.5)	163.0 (133.3-201.0)	<0.001
Organ failures n (%)
Circulatory failure	116 (27.8)	9 (4.5)	107 (49.1)	<0.001
Respiratory failure	131 (31.4)	7 (3.5)	124 (56.9)	<0.001
Cerebral failure	116 (27.8)	21 (10.6)	95 (43.6)	<0.001
Renal failure	87 (20.8)	0 (0)	120 (55.0)	<0.001
Coagulation failure	77 (18.4)	3 (1.5)	74 (33.9)	<0.001
Liver failure	62 (14.8)	4 (2.0)	58 (26.6)	<0.001
Specific treatments n (%)
Vasopressors	140 (33.5)	14 (7.0)	126 (57.8)	<0.001
Mechanical ventilation	131 (31.4)	7 (3.5)	124 (56.9)	<0.001
Renal replacement therapy	33 (7.9)	0 (0)	33 (15.1)	<0.001
Prognosis scores at enrollment
CTP	11 ( 8-12)	9 (7-11)	12 (10-13)	<0.001
MELD	19 ( 14-28)	14 (12-18)	28 (21-34)	<0.001
MELD-Na	22 ( 14-33)	15 (12-21)	31 (22-40)	<0.001
CTP class (A/B/C) n (%)	31 (7.4%）/123 (29.4%）/263 (63.2%）	30 (15.0%）/95 (47.7%）/74 (37.3%）	1 (0.4%）/28 (12.8%）/189 (86.8%）	<0.001
Length of ICU stay (days)	2 (1-4)	2 (1-2)	4 (2-6)	<0.001
Length of hospital stay (days)	7 (4-14)	5 (3-8)	12 (5-20)	<0.001
6-week mortality rate n (%)	124 (29.7)	22 (11.1)	102 (46.8)	<0.001

Abbreviations: ACLF, acute-on-chronic liver failure; HCC, hepatocellular carcinoma; HE, hepatic encephalopathy; SPO2, percutaneous oxygen saturation; FIO2, fraction of inspired oxygen; PVT, portal vein thrombosis; ALT, alanine aminotransferase; AST, aspartate aminotransferase; INR, international normalized ratio; CTP, Child-Turcotte-Pugh; MELD, model for end-stage liver disease; MELD-Na, MELD-sodium.

NOTE: data were described as means (±standard deviation), medians (interquartile range) or numbers (percentage) where appropriate and compared with independent sample t-test, Mann-Whitney U test and Chi square test accordingly.

Cumulative survival rates of cirrhotic patients hospitalized with acute variceal bleeding

Among all enrolled 417 patients, 124 patients were dead during a 42-day follow-up period and the cumulative survival rates at 42-days were 70.3%. A total of 102 patients with ACLF and 22 patients without ACLF were dead during follow-up, respectively. The cumulative 42-day survival rates in patients with ACLF was significantly lower than in those without ACLF (53.2% vs. 88.9%, P <0.001) (Fig. 2A). Furthermore, the cumulative 42-day survival rates decreased gradually as the grade of ACLF increased (P <0.001) (Fig. 2B) . A total of 9 (22.5%), 24 (34.2%) and 69 (63.8%) patients were dead among patients with ACLF grade I, II, and III during follow-up, respectively, (Fig. 2B).

ACLF as an independent risk factor of 6-week mortality in cirrhotic patients with AVB

To further validate the role of ACLF in the prognosis of cirrhotic patients hospitalized with AVB, we performed Cox-proportional hazards regression analysis based on the presence of ACLF or not at admission and the well recognized risk factors for the prognosis of cirrhotic patients with AVB. In the univariable analysis, the variables found to be statistically significant included race, mean arterial pressure (MAP), SPO2 (percutaneous oxygen saturation) / FIO2 (fraction of inspired oxygen), presence of ACLF at admission, concomitant of miocardial infarction and renal disease, hepatocellular carcinoma (HCC), ascites, hepatic encephalopathy (HE), white blood cell (WBC), hemoglobin (HGB), albumin (ALB), total bilirubin (TB), prothrombin time (PT), international normalized ratio (INR), alanine aminotransferase (ALT); aspartate aminotransferase (AST), serum sodium, serum potassium, serum creatinine (Scr), blood urea nitrogen (BUN) and the well established prognosis scores (Table 2). As INR and PT are both the biochemistry markers reflecting coagulation function, Scr and BUN reflecting renal function, we only select INR and Scr for further multivariable analysis. Multivariable analysis showed that only ACLF (HR, 2.74, 95% CI: 1.54-4.88, P <0.001), HCC (HR, 2.69, 95% CI: 1.64-4.43, P <0.001), MAP (HR, 0.96, 95% CI: 0.93-0.98, P <0.001), WBC (HR, 1.02, 95% CI: 1.00-1.04, P= 0.023), TB (HR, 1.03, 95% CI: 1.01-1.05, P <0.001), ALB (HR, 0.55, 95% CI: 0.40-0.76, P <0.001) and INR (HR, 1.24, 95% CI: 1.10-1.39, P <0.001) remained in the final regression model (Fig. 2).

Table 2 Risk factors for 6-week mortality in all the cirrhotic patients hospitalized with AVB (N = 417)

Variable	univariable analysis HR (95% CI) P value	multivariable analysis HR (95% CI) P value
Gender
Female	0.67 (0.44-1.01) 0.057	0.73 (0.47-1.15) 0.179
Age	1.01 (0.99-1.02) 0.426
Race
Non white	1.63 (1.14-2.33) 0.007	1.11 (0.74-1.66) 0.614
Etiology of cirrhosis
Non alcohol	0.97 (0.68-1.38) 0.845
Concurrent HCC	1.66 (1.05-2.63) 0.031	2.69 (1.64-4.43) <0.001
Infection	1.32 (0.91-1.93) 0.149
Ascites	1.99 (1.39-2.95) 0.001	0.99 (0.64-1.55) 0.981
Hepatic encephalopathy	1.37 (1.10-1.71) 0.005	0.81 (0.62-1.06) 0.124
PVT	1.55 (0.97-2.48) 0.068
Mean arterial pressure	0.94 (0.92-0.96) <0.001	0.96 (0.93-0.98) <0.001
SPO2/FIO2	1.00 (1.00-1.00) 0.001	1.00 (1.00-1.00) 0.641
White blood cell	1.04 (1.03-1.05) <0.001	1.02 (1.00-1.04) 0.023
Hemoglobin	0.82 (0.74-0.91) <0.001	0.91 (0.82-1.01) 0.063
Platelet	1.00 (0.99-1.00) 0.230
Total bilirubin	1.04 (1.03-1.05) <0.001	1.03 (1.01-1.05) <0.001
Albumin	0.39 (0.29-0.54) <0.001	0.55 (0.40-0.76) <0.001
ALT	1.00 (1.00-1.00) 0.002	1.00 (1.00-1.00) 0.387
AST	1.00 (1.00-1.00) 0.009
INR	1.43 (1.33-1.53) <0.001	1.24 (1.10-1.39) <0.001
Prothrombin time	1.02 (1.02-1.03) <0.001
Serum creatinine	1.19 (1.13-1.26) <0.001	1.00 (0.90-1.10) 0.923
Blood urea nitrogen	1.01 (1.00-1.01) 0.001
Serum sodium	0.97 (0.94-0.99) 0.016	1.02 (0.99-1.05) 0.185
Serum potassium	1.46 (1.23-1.72) <0.001	1.16 (0.95-1.41) 0.157
Glucose	1.00 (1.00-1.00) 0.229
ACLF	5.52 (3.48-8.76) <0.001	2.74 (1.54-4.88) <0.001
Concomitant disease
Miocardial infarction	2.48 (1.31-4.74) 0.006	1.22 (0.56-2.67) 0.62
Congestive heart failure	1.04 (0.54-1.98) 0.909
Cerebrovascular disease	1.62 (0.76-3.48) 0.213
Chronic pulmonary disease	0.88 (0.54-1.46) 0.625
Diabetes	1.50 (0.83-2.72) 0.181
Peripheral vascular disease	1.78 (0.78-4.04) 0.169
Renal disease	1.74 (1.09-2.79) 0.020	1.21 (0.73-2.01) 0.464
Prognosis scores
CTP	1.38 (1.17-1.62) <0.001
MELD	1.10 (1.08-1.12) <0.001
MELD-Na	1.06 (1.05-1.08) <0.001

Abbreviations: ACLF, acute-on-chronic liver failure; SPO2, percutaneous oxygen saturation; FIO2, fraction of inspired oxygen; HCC, hepatocellular carcinoma; PVT, portal vein thrombosis; ALT, alanine aminotransferase; AST, aspartate aminotransferase; INR, international normalized ratio; CTP, Child-Turcotte-Pugh; MELD, model for end-stage liver disease. MELD-Na, MELD-sodium.

Prediction performance of prognosis scores

1 Discrimination performance

To validate the discrimination performance of the common prognosis scores for the 6-week mortality in cirrhotic patients hospitalized with AVB, we performed ROC on them in ACLF (Fig. 3A) and AD patients (Fig. 3B), respectively. In the ACLF patients, the AUROC calculated for CTP, MELD, MELD-Na were 0.658 (95% CI, 0.591–0.721), 0.728 (95% CI, 0.664–0.786), 0.725 (95% CI, 0.660–0.783) and 0.729 (95% CI, 0.665–0.787), respectively (Table 3). In the Delong test, there were no significant difference among the AUROC for these prognosis scores (P >0.05). The cutoff value, sensitivity and specificity for above mentioned prognosis scores were 11/73.5/53.5, 25/80.4/58.8, 26/88.2/52.6 and 61/67.7/73.3, respectively (Table 3). In the AD patients, the AUROC calculated for CTP, MELD, MELD-Na and CLIF-C AD were 0.686 (95% CI, 0.617–0.750), 0.674 (95% CI, 0.605–0.739), 0.667 (95% CI, 0.597–0.732) and 0.737 (95% CI, 0.670–0.797), respectively (Table 3). In the Delong test, there were no significant difference among the AUROC for these prognosis scores (P >0.05). The cutoff value, sensitivity and specificity for above mentioned prognosis scores were 10/54.6/77.4, 14/72.7/55.4, 23/40.9/88.1 and 55/77.3/68.4, respectively (Table 3).

To validate the calibration performance of the common prognosis scores for the 6-week mortality in cirrhotic patients hospitalized with AVB, we performed Hosmer-Lemeshow goodness-of-fit test for these prognosis scores (Table 3). In the ACLF patients, the P value between observed and predicted probability of 6-week mortality for CTP, MELD, MELD-Na and CLIF-C ACLF were 0.024, 0.650, 0.004 and 0.491, respectively. In the AD patients, the P value between observed and predicted probability of 6-week mortality for CTP, MELD, MELD-Na and CLIF-C ADs were 0.836, 0.670, 0.554 and 0.526, respectively.

To overcome the shortcomings of Hosmer-Lemeshow goodness-of-fit test in the evaluation of calibration, we plotted the calibration curves for the prognostic scores (Fig. 4). In ACLF patients, the visually inspected concordance between observed and predicted probability of 6-week mortality for CLIF-C ACLF were excellent and superior to those for the other prognosis scores (Fig. 4A). In AD patients, the visually inspected concordance between observed and predicted probability of 6-week mortality for MELD, MELD-Na and CLIF-C ADs were satisfactory and superior to that for CTP (Fig. 4B).

3 Overall performance

To validate the overall performance of the common prognosis scores for the 6-week mortality in cirrhotic patients hospitalized with AVB, we calculated the Brier score and R²valuefor these prognosis scores (Table 3). In the ACLF patients, the Brier score and R²value for CTP, MELD, MELD-Na and CLIF-C ACLF were 0.229/0.104, 0.210/0.203, 0.213/0.187 and 0.209/0.219, respectively. In the AD patients, the Brier score and R²value for CTP, MELD, MELD-Na and CLIF-C ADs were 0.094/0.077, 0.094/0.068, 0.093/0.073 and 0.087/0.173, respectively.

Table 3 Predictive values of prognostic scores

Scores	AUC	Youden index	Cutoff value	SEN	SPE	PPV	NPV	Brier	R²	P in H-L test
ACLF patients
CTP	0.658	0.27	11	73.5	53.5	58.1	69.7	0.229	0.104	0.024
MELD	0.728	0.39	25	80.4	58.8	63.1	77.3	0.210	0.203	0.650
MELD-Na	0.725	0.41	26	88.2	52.6	62.1	83.6	0.213	0.187	0.004
CLIF-C ACLF	0.729	0.41	61	67.7	73.3	69.0	72.0	0.209	0.219	0.491
AD patients
CTP	0.686	0.32	10	54.6	77.4	23.1	93.2	0.094	0.077	0.836
MELD	0.674	0.28	14	72.7	55.4	16.8	94.2	0.094	0.068	0.670
MELD-Na	0.667	0.29	23	40.9	88.1	30.0	92.3	0.093	0.073	0.554
CLIF-C AD	0.737	0.46	55	77.3	68.4	23.3	96.0	0.087	0.149	0.526

Abbreviations: AUC, area under receiver operating characteristic curve; SEN, sensitivity; SPE, specificity; PPV, positive predictive value; NPV, negative predictive value; H-L test, Hosmer-Lemeshow goodness-of-fit test. CLIF-C ACLF, Chronic liver failure-organ failure-Consortium acute-on-chronic liver failure; CLIF-C AD, Chronic liver failure-organ failure-Consortium Acute Decompensation.

Acute variceal bleeding (AVB) is a life-threatening complication of cirrhosis and accurate risk stratification of AVB is essential for clinicians to provide rational treatment. Acute-on-chronic liver failure (ACLF) is a syndrome characterized by acute decompensation of cirrhosis with organ failures and high short-term mortality. In this single center retrospective cohort study, we investigated the role of ACLF in the prognosis of cirrhotic patients with AVB and found that the cumulative 6-week survival rate in patients with ACLF was significantly lower than that in those without ACLF and as the ACLF grades increased, the 6-week survival rate significantly decreased. In addition, the presence of ACLF at admission remained as a risk factor for 6-week mortality in cirrhotic patients with AVB after adjusting for confounding factors. Compared with other common prognosis scores, CLIF-C ACLF and CLIF-C AD performed best in patients with and without ACLF, respectively. To the best of our knowledge, this was the first study to validate the prediction performance of prognosis scores in AVB patients with and without ACLF, respectively.

In the last decade, different definitions of ACLF have been developed by multiple international consortia ^{[17][25][26][27]}. Among them, we choose the EASL-CLIF definition ^[17] because it was most suitable for the design of this study. The diagnosis of ACLF is of great importance because it would allow early identification of patients at high risk for end-organ failure–related death, requiring specific treatments and/or intensive management ^[17]. In this study, patients with ACLF more frequently presented with other complications of cirrhosis (ascites, bacterial infections and hepatic encephalopathy), various organ failures and related abnormal biochemistry markers as well as significantly higher peripheral white blood cell (WBC) count and lower serum sodium. It was not surprising that patients with ACLF more frequently presented with ascites and bacterial infections because they are both independent predictive factors of kidney failure [^28][29]. Current evidence shows that the pathophysiology of ACLF is closely associated with an intense systemic inflammation, as indicated by a high WBC count, which may induce immune-mediated tissue damage and mitochondrial dysfunction contributing to the development of organ failures [^30][31][32]. Hyponatremia has been well described in associations with hepatorenal syndrome, ascites and liver-related mortality ^[14]and therefore maybe also involved in the development of ACLF.

Although it is well recognized that cirrhotic patients with ACLF carry a high short-term mortality, there are limited literature about the role of ACLF in the prognosis of cirrhotic patients with AVB, exclusively [^33][34][35]. In this study, the 6-week mortality in patients with ACLF was significantly higher than those without ACLF (46.8% vs. 11.1%, P <0.0001) and increased significantly as the grade of ACLF increased (grade 1: 22.5%, grade 2: 34.2% and grade 3: 63.8%, P <0.0001) (Fig 1). This was consistent with the result in the study by Trebicka et al (6-week mortality 47.1% vs. 10.0%, grade 1: 30.0%, grade 2: 50.0% and grade 3: 70.0%) ^[33], Kumar et al (6-week mortality 47.9% vs. 9.1%, grade 1: 24.0%, grade 2: 44.0% and grade 3: 77.0%) ^[34]and Shin et al (28-day mortality 41.0% vs. 3.4%, grade 1: 7.1%, grade 2: 28.6% and grade 3: 80.8%) ^[35]. The different mortality rate in the subgroups of ACLF among these studies might be explained by the different baseline characteristics of enrolled patients or different diagnostic criteria of ACLF. For example, patients with hepatocellular carcinoma, severe chronic extra-hepatic disease and chronic decompensation of end-stage liver disease (ascites, hepatic encephalopathy and infection) were all excluded from the study by Shin et al and CLIF-SOFA score was used for the diagnosis of ACLF. In addition, our mortality was consistent with that in the CANONIC study ^[17] (28-day and 90-day mortality in cirrhotic patients without ACLF, with ACLF, ACLF grade 1, grade 2 and grade 3 were 4.7%/14%, 33.9%/51.2%, 22.1%/40.7%, 32.0%/52.3% and 76.7%/79.1%) where the definition and diagnostic criteria of EASL-ACLF was firstly proposed by EASL-CLIF Consortium, except that in ACLF grade 3. It might be explained by the different patients included and diagnostic criteria of ACLF between the two studies, that is, the CANONIC study enrolled cirrhotic patients with various acute decompensations and diagnosed ACLF according to CLIF-SOFA criteria, whereas we only included cirrhotic patients with AVB and diagnosed ACLF with CLIF-OF criteria. In multivariate analysis, the presence of ACLF remained as an independent risk factor for the 6-week mortality after adjusting for confounding factors (HR =2.74, CI: 1.54-4.88, p <0.001). This was highly consistent with the finding in the study by Trebicka et al (HR =2.72, CI: 1.95-3.78, P <0.001), which to some extent indicates that our result was highly credible.

Although numerous studies have validated the prediction performance of prognosis scores in the setting of cirrhosis with AVB, few have validated that in patients with or without ACLF which plays a significant role in the prognosis of these patients. In this study, we validated the prediction performance of common used prognosis scores in patients with and without ACLF (mere AD) along the three metrics (discrimination, calibration and overall performance). In ACLF patients, although without significant difference of the AUROC among the 4 prognosis scores, the calibration performance of CLIF-C ACLF are the best (Fig. 4A). In addition, the overall performance of CLIF-C ACLF was superior to the other prognosis scores (with the lowest Brier score and highest R² value). In AD patients, CLIF-C AD was the only prognosis score with AUROC >0.7 (although without significant difference when compared to that for the other prognosis scores). The calibration performance of MELD, MELD-Na and CLIF-C AD are comparable and superior to that for CTP (Fig. 4B). In addition, the overall performance of CLIF-C AD was superior to the other prognosis scores (with the lowest Brier score and highest R² value). Furthermore, we performed sensitivity analysis and the results showed that the sensitivity and specificity of CLIF-C ACLF and CLIF-C AD were satisfactory and superior to that of the other prognosis scores (table 3). The superior prediction performance of CLIF-C ACLF and CLIF-C AD to the other prognosis scores in ACLF and AD patients might be due to the distinct background where these prognosis scores were developed. CTP and MELD were developed in cirrhotic patients who received surgery and TIPS therapy due to recurrent esophagogastric variceal bleeding and various complications of portal hypertension, respectively. MELD-Na was developed in cirrhotic patients listed for liver transplantation. All these three prognosis scores were developed regardless of the presence of ACLF. On the other hand, CLIF-C ACLF and CLIF-C AD were developed by EASL-CLIF (European Association for the Study of Liver-Chronic Liver Failure) Consortium in patients included in the CANONIC study^[16]with and without ACLF, respectively. Thus, a superior prediction performance of these two prognosis scores could be expected. According to the cutoff value of CLIF-C ACLF and CLIF-C AD, ACLF patients with CLIF-C ACLF score >61 and AD patients with CLIF-C ADs >55 may need to be stratified as high risk, respectively. These patients may need to be provided with organ support in intensive care unit ^[24] or salvage treatment such as pre-emptive TIPS ^[33]or Rescue TIPS ^[34], which were recently found to be very effective to reduce the 6-week and 1 year mortality in AVB patients with ACLF.

Our study has some limitations. First, as this was a single center and observational study, selection, information and confounding biases were inevitable. Second, although the medical record information in the MIMIC database was prospectively and timely collected, the diagnosis of covert hepatic encephalopathy (minimal hepatic encephalopathy and grade Ⅰ ) might be partly effected by the subjective factors of observers. Besides, the diagnosis of respiratory failure or brain failure became difficult when the mechanical ventilation was provided because it was hard to determine the exact reason for the mechanical ventilation therapy (respiratory failure, brain failure or both), which might to some extent lead to a bias in the diagnosis of respiratory failure or brain failure. Nonetheless, since the mortality in our cohort was consistent with that described in the recent prospective studies about ACLF ^{[33][34][35][36]}, we believe that our results are highly credible. Finally, since the enrolled patients were limited to those hospitalized in the ICU and did not include those in general ward, we did not validate the prediction performance of prognosis scores in all cirrhotic patients with AVB who visited the Beth Israel Deaconess Medical Center.

In conclusion, ACLF at admission is an independent predictor for the 6-week mortality in cirrhotic patients with AVB. To improve the prognosis of these patients, it is essential to formulate therapeutic schedule and inform the patient's family about the prognosis according to the presence of ACLF or not at admission. For patients with and without ACLF, CLIF-C ACLF and CLIF-C AD outperformed other prognostic scores in the prediction of 6-week mortality, respectively, and can be used for the risk stratification of these two distinct groups of patients.

Acknowledgements

We would like to thank the researchers at Massachusetts Institute of Technology Laboratory for Computational Physiology and collaborating research groups for their Development and establishment of the MIMIC database and thank the medical staff in the Beth Israel Deaconess Medical Center for their continued support of the MIMIC project.

Author Contributions

Zongyi Zhu collected and analyzed the data and wrote the paper. Huiqing Jiang designed the study and revised the paper.

Publication approval

All authors approved the publication of this article.

Conflicts of interest

None declared

Data link

https://www.scidb.cn/s/uqmYNf.

Source of funding

This study was supported by grant from Health Care and Biomedicine Special Project Hebei Province Key R&D Program (182777117D).

Reverter E, Tandon P, Augustin S, Turon F, Casu S, Bastiampillai R, et al. A MELD-based model to determine risk of mortality among patients with acute variceal bleeding. Gastroenterology 2014;146:412-419.
Fortune B, Garcia-Tsao G, Ciarleglio M, Deng Y, Fallon MB, Sigal S, et al.Vapreotide Study Group. Child-Turcotte-Pugh Class is best at stratifying risk in variceal hemorrhage: analysis of a U.S. multi-center prospective study. J Clin Gastroenterol 2016 Oct 24.
European Association for the Study of the Liver. EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis.J Hepatol.2018 Aug;69(2):406-460.
de Franchis R. Expanding consensus in portal hypertension: Report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension. J Hepatol. 2015;63(3):743–52.
Garcia-Tsao G, Abraldes JG, Berzigotti A, Bosch J. Portal hypertensive bleeding in cirrhosis: Risk stratification, diagnosis, and management: 2016 practice guidance by the American Association for the study of liver diseases. Hepatology. 2017;65(1):310–35.
Tripathi D, Stanley AJ, Hayes PC, Patch D, Millson C, Mehrzad H, et al. U.K. guidelines on the management of variceal hemorrhage in cirrhotic patients. Gut. 2015;64(11):1680–704.
de Franchis R. Expanding consensus in portal hypertension: Report of the Baveno VI Consensus Workshop: Stratifying risk and individualizing care for portal hypertension. J Hepatol 2015;63(3): 743–752.
Garcia-Tsao G, Abraldes JG, Berzigotti A, Bosch J. Portal hypertensive bleeding in cirrhosis: Risk stratification, diagnosis, and management: 2016 practice guidance by the American Association for the study of liver diseases. Hepatology 2017;65(1): 310–335.
Tripathi D, Stanley AJ, Hayes PC, Patch D, Millson C, Mehrzad H, et al. U.K. guidelines on the management of variceal hemorrhage in cirrhotic patients. Gut 2015;64(11): 1680–704.
Salvador Augustin, Laura Muntaner, José T Altamirano, Antonio González, Esteban Saperas, Joan Dot, et al. Predicting early mortality after acute variceal hemorrhage based on classification and regression tree analysis. Clin Gastroenterol Hepatol 2009;7(12): 1347-1354.
Ripoll C, Banares R, Rincon D, Catalina MV, Lo Iacono O, Salcedo M, et al. Influence of hepatic venous pressure gradient on the prediction of survival of patients with cirrhosis in the MELD Era. Hepatology 2005;42: 793-801.
Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60: 646–649.
Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter Borg PCJ. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology. 2000; 31: 864–71.
Scott W Biggins, W Ray Kim, Norah A Terrault, Sammy Saab, Vijay Balan, Thomas Schiano, et al. Evidence-based incorporation of serum sodium concentration into MELD. Gastroenterology 2006;130(6): 1652-60.
Ruben Hernaez, Elsa Solà, Richard Moreau, Pere Ginès.Acute-on-chronic liver failure: an update.Gut. 2017 Mar;66(3):541-553.
Sundaram V, Jalan R, Wu T, et al. Factors associated with survival of patients with severe acute-on-chronic liver failure before and after liver transplantation. Gastroenterology 2019;156:1381– 91.e3.
Moreau R. Jalan R.Gines P.Pavesi M.Angeli P.Cordoba J.et al. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology. 2013; 144 (1437.e1–9): 1426-1437
Garg H, Kumar A, Garg V, Kumar M, Kumar R, Sharma B.C. et al. Hepatic and systemic hemodynamic derangements predict early mortality and recovery in patients with acute-on-chronic liver failure. J Gastroenterol Hepatol. 2013; 28: 1361-1367
Trebicka J. Emergency TIPS in a Child-Pugh B patient: when does the window of opportunity open and close?.J Hepatol. 2017; 66: 442-450
Jonel Trebicka, Javier Fernandez, Maria Papp, Paolo Caraceni, Wim Laleman, Carmine Gambino, et al. PREDICT identifies precipitating events associated with the clinical course of acutely decompensated cirrhosis. J Hepatol. 2021 May;74(5):1097-1108.
Johnson, A., Bulgarelli, L., Pollard, T., Horng, S., Celi, L. A., & Mark, R. (2022). MIMIC-IV (version 2.0). PhysioNet. https://doi.org/10.13026/7vcr-e114.
Goldberger, A., Amaral, L., Glass, L., Hausdorff, J., Ivanov, P. C., Mark, R., ... & Stanley, H. E. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation [Online]. 101 (23), pp. e215–e220.
Jalan R, Saliba F, Pavesi M, Amoros M, Moreau R, Gines P, et al. Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure. J Hepatol 2014; 61: 1038-1047.
Jalan R, Pavesi M, Saliba F, Amorós A, Fernandez J, Holland-Fischer P, et al. The CLIF Consortium Acute Decompensation score (CLIF-C ADs) for prognosis of hospitalised cirrhotic patients without acute-on-chronic liver failure. J Hepatol 2015;62: 831–840.
Bajaj J.S., O'Leary J.G., Reddy K.R., Wong F., Biggins S.W., Patton H. Survival in infection-related acute-on-chronic liver failure is defined by extrahepatic organ failures. Hepatology. 2014;60:250–256.
Wu T., Li J., Shao L., Xin J., Jiang L., Zhou Q., on behalf of the Chinese Group on the Study of Severe Hepatitis B (COSSH) Development of diagnostic criteria and a prognostic score for hepatitis B virus-related acute-on-chronic liver failure. Gut. 2018;67:2181–2191.
Sarin S.K., Choudhury A., Sharma M.K., Maiwall R., Al Mahtab M., Rahman S. Acute-on-chronic liver failure: consensus recommendations of the Asian Pacific Association for the Study of the Liver (APASL): an update. Hepatol Int. 2019;13:353–390.
Moreau R. Lebrec D. Acute renal failure in patients with cirrhosis: perspectives in the age of MELD.Hepatology. 2003; 37: 233-243
Ginès P. Schrier R.W. Renal failure in cirrhosis.N Engl J Med. 2009; 361: 1279-1290
Rahul Kumar, Gautam Mehta, Rajiv Jalan. Acute-on-chronic liver failure. Clin Med (Lo. 2020;20(5): 501-504.
Medzhitov R, Schneider D.S, Soares M.P. Disease tolerance as a defense strategy. Science. 2012; 335: 936-941
Moreau R. Clària J. Aguilar F. Fenaille F. Lozano J.J. Junt C. et al. Blood metabolomics uncovers inflammation-associated mitochondrial dysfunction as a potential mechanism underlying ACLF. J Hepatol. 2020; 72: 688-701
Jonel Trebicka , Wenyi Gu , Luis Ibáñez-Samaniego , Virginia Hernández-Gea , Carla Pitarch, et al. Rebleeding and mortality risk are increased by ACLF but reduced by pre-emptive TIPS. J Hepatol 2020;73(5): 1082-1091.
Rahul Kumar, Annarein J C Kerbert, M Faisal Sheikh, Noam Roth, Joana A F Calvao, Monica D Mesquita, et al. Determinants of mortality in patients with cirrhosis and uncontrolled variceal bleeding. J Hepatol. 2021;74(1): 66-79.
Jongbeom Shin, Jung Hwan Yu, Young-Joo Jin, Hyung Joon Yim, Young Kul Jung, et al. Acute-on-chronic liver failure as a major predictive factor for mortality in patients with variceal bleeding. Clin Mol Hepatol 2020; 26(4): 540-553.
Zhujun Cao, Yuhan Liu, Minghao Cai, Yumin Xu, Xiaogang Xiang, Gangde Zhao, et al. The Use of NACSELD and EASL-CLIF Classification Systems of ACLF in the Prediction of Prognosis in Hospitalized Patients With Cirrhosis. Am J Gastroenterol. 2020 ;115(12):2026-2035.

Risk Stratification Based on Acute-on-chronic liver failure in cirrhotic patients hospitalized with acute variceal bleeding

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Patients And Methods

Patients

Data collection

Study outcomes

Statistical analysis

Results

Prediction performance of prognosis scores

Discussion

Declarations

References

Status:

Journal Publication

Version 1