Next Article in Journal
Sudden Vision Loss Due to Optic Neuritis—An Uncommon Presentation of Neurosarcoidosis
Next Article in Special Issue
Diagnosis, Management, and Prognosis of Cystic Fibrosis-Related Liver Disease in Children
Previous Article in Journal
Breast Self-Examination Practice and Its Determinants among Women in Indonesia: A Systematic Review, Meta-Analysis, and Meta-Regression
Previous Article in Special Issue
Hemoglobin and Endotoxin Levels Predict Sarcopenia Occurrence in Patients with Alcoholic Cirrhosis
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Diagnostics
Volume 13
Issue 15
10.3390/diagnostics13152578
diagnostics-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Validated Composite Score Demonstrates Potential Superiority to MELD-Based Systems in Predicting Short-Term Survival in Patients with Liver Cirrhosis and Spontaneous Bacterial Peritonitis—A Preliminary Study

by
Yan-Ting Lin
1,
Wei-Ting Chen
1,2,
Tsung-Han Wu
2,3,
Yu Liu
1,
Li-Tong Liu
1,
Wei Teng
1,2,
Yi-Chung Hsieh
1,2,
Yen-Mu Wu
2,4,
Chien-Hao Huang
1,2,*,
Chao-Wei Hsu
1,2 and
Rong-Nan Chien
1,2
1
Department of Hepatology, Department of Gastroenterology and Hepatology, Chang-Gung Memorial Hospital, Linkou Medical Center, Taoyuan 33305, Taiwan
2
College of Medicine, Chang-Gung University, Taoyuan 33302, Taiwan
3
Department of General Surgery, Chang-Gung Memorial Hospital, Linkou Medical Center, Taoyuan 333, Taiwan
4
Department of Infectious Disease, Chang-Gung Memorial Hospital, Linkou Medical Center, Taoyuan 33305, Taiwan
*
Author to whom correspondence should be addressed.
Diagnostics 2023, 13(15), 2578; https://doi.org/10.3390/diagnostics13152578
Submission received: 15 June 2023 / Revised: 22 July 2023 / Accepted: 25 July 2023 / Published: 2 August 2023
(This article belongs to the Special Issue Diagnosis, Management, and Prognosis of Liver Disease and Its Syndromes)
Browse Figures
Versions Notes

Abstract

:
Background: Spontaneous bacterial peritonitis (SBP) is a severe complication in cirrhosis patients with ascites, leading to high mortality rates if not promptly treated. However, specific prediction models for SBP are lacking. Aims: This study aimed to compare commonly used cirrhotic prediction models (CTP score, MELD, MELD-Na, iMELD, and MELD 3.0) for short-term mortality prediction and develop a novel model to improve mortality prediction. Methods: Patients with the first episode of SBP were included. Prognostic values for mortality were assessed using AUROC analysis. A novel prediction model was developed and validated. Results: In total, 327 SBP patients were analyzed, with HBV infection as the main etiologies. MELD 3.0 demonstrated the highest AUROC among the traditional models. The novel model, incorporating HRS, exhibited superior predictive accuracy for in-hospital in all patients and 3-month mortality in HBV-cirrhosis, with AUROC values of 0.827 and 0.813 respectively, surpassing 0.8. Conclusions: MELD 3.0 score outperformed the CTP score and showed a non-significant improvement compared to other MELD-based scores, while the novel SBP model demonstrated impressive accuracy. Internal validation and an HBV-related cirrhosis subgroup sensitivity analysis supported these findings, highlighting the need for a specific prognostic model for SBP and the importance of preventing HRS development to improve SBP prognosis.

Graphical Abstract

1. Introduction

Patients with liver cirrhosis often exhibit an impaired defense against bacteria, leading to reduced bacterial clearance [1]. This immune defect, coupled with increased intestinal permeability and gut bacterial overgrowth, facilitates bacterial translocation. As a result of the pathological bacterial translocation in liver cirrhosis, the bacterial infection is either present on admission or develops during hospitalization in approximately 30% of patients with cirrhosis [2]. The most prevalent form of these infections is spontaneous bacterial peritonitis (SBP) [2].
SBP is characterized by the spontaneous infection of ascitic fluid, occurring in cirrhotic patients in the absence of an intra-abdominal source of infection [3]. The presence of SBP in patients with ascites can lead to high mortality rates if not promptly treated [4,5]. Additionally, the concurrence of hepatorenal syndrome (HRS), characterized by acute kidney injury in patients with acute or chronic liver disease and affecting up to 30% of patients with SBP, further increases the risk of mortality [3]. One-year overall mortality rates for SBP range from 53.9% to 78% [6,7], indicating that SBP may represent the final clinical stage of liver cirrhosis [8]. Given the serious consequences of SBP, accurate prediction of mortality is crucial for guiding clinical decisions and considering liver transplantation. However, it is worth noting that specific prediction models for SBP are currently lacking [5].
The model for end-stage liver disease (MELD) has demonstrated superior accuracy in predicting 3-month mortality compared to the Child–Turcotte–Pugh (CTP) score in cirrhotic patients awaiting liver transplantation in the United States [9]. Therefore, the MELD has been widely adopted by the United Network for Organ Sharing (UNOS) and other countries since 2002 to improve the prioritization of liver transplantation (LT) waiting lists [10]. Several MELD-based models, such as the model for end-stage liver disease with the incorporation of serum sodium (MELD-Na) [11] and the integrated model for end-stage liver disease (iMELD) score [12], have been developed to enhance mortality prediction in cirrhotic patients [13,14]. In recent years, MELD has been updated to version 3.0 to address the need for reducing deaths while patients are on the waitlist for liver transplantation [15]. Despite the good prognostic performance of iMELD and MELD-Na for 3-month and 6-month mortality in cirrhosis patients [16,17], their applicability in subgroups with liver cirrhosis-related complications, such as SBP, remains limited [18,19]. This limitation is particularly relevant in regions with a high prevalence of chronic hepatitis B, such as Taiwan, where chronic hepatitis B virus (HBV) infection is the leading cause of HCC, especially in Asia and Africa [20,21].
Currently, there is a lack of specific prediction models for SBP. However, in our previous study, we demonstrated that among the MELD-based models, iMELD and MELD-Na exhibited the highest area under the receiver operating characteristic curve (AUROC) and significantly outperformed other models in predicting 3-month and 6-month mortalities in patients with HBV-related liver cirrhosis and SBP [22]. Subsequently, we also identified that factors such as bacteremia (sepsis), hepatorenal syndrome (HRS), and serum creatinine levels had a stronger impact on short-term survival compared to types of SBP or microbial patterns in these patients [23]. Therefore, the objectives of our study are to (i) compare commonly used cirrhotic mortality models, including the recently updated MELD 3.0 [15], and (ii) explore the potential for developing a superior prediction model to accurately predict short-term mortality (in-hospital, 3 and 6 month) in patients with SBP.

2. Materials and Methods

2.1. Patient Selection

A retrospective cohort study was conducted with approval from the ethical committees of Chang Gung Memorial Hospital. As shown in Figure 1, adult patients diagnosed with SBP and liver cirrhosis between January 2006 and August 2017 were included in the study. Patient data, including clinical, demographic, hematological, and biochemical information, were collected at the time of SBP diagnosis. Various scoring systems, such as the CTP score and the four MELD-based scores (MELD, MELD-Na, iMELD, MELD 3.0), were calculated. Patients were followed up until March 2018 or until death. Patients who had previous episodes of SBP, had malignancy, had ascites not related to cirrhosis, or underwent liver transplantation within 6 months of follow-ups were excluded from mortality calculations. Consequently, a total of 327 patients diagnosed with the first episode of SBP were enrolled in the analysis.

2.2. Diagnosis, Definition, and Management of Liver Cirrhosis with Ascites and Spontaneous Bacterial Peritonitis

The diagnosis of liver cirrhosis (LC) with ascites was based mainly on the following criteria: (1) LC diagnosed based on histopathology (liver biopsy) or compatible clinical features, laboratory data, sonographic/computed tomography/magnetic resonance imaging findings typical for liver cirrhosis [24], or non-invasive testing such as FibroScan [25]; (2) ascites caused by portal hypertension (serum-ascites albumin gradient (SAAG) ≥ 1.1 g/dL) [26]; (3) exclusion of other underlying disease that can cause ascites such as malignancy (HCC or metastasis), right-sided congestive heart failure/constrictive pericarditis, Budd–Chiari syndrome, post-sinusoidal obstruction syndrome, portal or splenic vein thrombosis, and the possibility of schistosomiasis.
SBP is suspected in cirrhotic patients presenting with suggestive symptoms and signs, such as fever, abdominal pain, altered mental status, abdominal tenderness, or hypotension [27]. SBP is diagnosed upon an absolute neutrophil count in ascites fluid of ≥250 cells/mm3 and/or positive ascites culture, in the absence of a surgically treatable intra-abdominal source of infection and other causes of elevated ascites neutrophil count, such as hemorrhage, pancreatitis, peritoneal tuberculosis, or carcinomatosis [28]. Since infections are common in cirrhotic patients with variceal bleeding [29], paracentesis should be performed upon hospital admission in all cirrhotic patients with ascites, even if admitted for reasons other than ascites [30]. Baseline laboratory data were collected when SBP was suspected, and prompt diagnosis was facilitated by efficient laboratory services. In our facility, the results of ascites fluid routine can be obtained within 2 h, enabling a timely diagnosis of SBP within 1 day.
The management of SBP follows the recommendations of the International Ascites Club [31] and the AASLD/EASL/JAMA guidelines [5,25]. In the case of HBV-related cirrhotic patients, antiviral agents are administered if they meet the latest APASL guideline [32,33,34].

2.3. Primary Outcomes and Scheduled Follow-Up Periods

Due to the high short-term mortality associated with SBP, the primary outcome of this study was defined as in-hospital mortality, as well as mortality at 3 months (3 M) and 6 months (6 M) following diagnosis. Each patient was followed up at least every 3 to 6 months until the date of death or March 2018, whichever occurred first.

2.4. Calculation of the Child–Turcotte–Pugh, MELD, MELD-Na, iMELD, and MELD 3.0 Scores

The Child–Turcotte–Pugh (CTP) score was calculated according to the established criteria [35]. MELD, MELD-Na, iMELD, and MELD 3.0 scores were measured using the formula proposed in previous studies. The MELD score was 11.2 × loge (INR) + 9.57 × loge (creatinine, mg/dL) + 3.78 × loge (bilirubin, mg/d) + 6.43), with a lower bound of 1 for all 3 variables and an upper bound of 4 for serum creatinine [9]. The MELD-Na score was MELD-Na − [0.025 × MELD × (140 − Na)] + 140, with Na was bounded at 125 and 140 [11]. The iMELD score was MELD + (Age × 0.3) − (0.7 × Na + 100) [12]. The MELD 3.0 was 1.33 × (Female) + 4.56 × ln (Serum bilirubin) + 0.82 × (137 − Sodium) − 0.24 × (137 − Sodium) × ln (Serum bilirubin) + 9.09 × ln (INR) + 11.14 × ln (Serum creatinine) + 1.85 × (3.5 − Serum albumin) − 1.83 × (3.5 − Serum albumin) × ln (Serum creatinine) + 6, rounded to the nearest integer [15].

2.5. Statistical Analysis

The data were properly tabulated for statistical analysis. Normally distributed continuous data were presented as mean ± standard deviation (SD) and compared using an independent Student’s t-test. Non-normally distributed continuous data were expressed as median and interquartile range (IQR) and compared using the Mann–Whitney U test. Categorical variables were presented as frequencies and percentages and compared using the χ2 test or Fisher’s exact test. The predictive ability of the novel model was evaluated by calculating the area under the receiver operating characteristic curve (AUROC) and compared to existing scoring systems such as CTP, MELD, MELD-Na, iMELD, and MELD 3.0. The predictive performance of each scoring system in predicting mortality was compared using the DeLong test. The best cut-off values for the new SBP scoring system were determined using the Youden Index. All statistical analyses were performed using IBM SPSS Statistics version 25 (SPSS Inc., Chicago, IL, USA) and R 4.2.2 (R Core Team, 2021; Vienna, Austria). A p-value less than 0.05 was considered statistically significant.

2.6. Construction of a Novel Prognostic Model

Using the current dataset, the primary outcomes were the dependent variable in the construction of a novel prognostic model. Initially, several independent variables were included, such as patients’ demographics, etiology of cirrhosis, ascites PMN, the presence or absence of sepsis, ascites, hepatic encephalopathy (HE), HRS, serum Cr, Na, K, total bilirubin, albumin, INR, hemoglobin, WBC, and PLT. Variables with a p < 0.20 in the univariate analysis were included in a multivariate logistic regression analysis [36] to identify independent predictors that could predict primary outcomes with an AUROC > 0.8, indicating a good discriminatory ability [37].

2.7. Model Validation and Sensitivity Analysis

For model validation, internal validation was performed using 5-fold cross-validation, which involved dividing a group of 327 patients into 5 equal subsets. The model was trained 5 times, with each iteration using 4 out of the 5 subsets as the training data (261 patients), while the remaining subset (66 patients) served as the test (validation) set. The performance of the model was assessed using various metrics, including AUROC, Brier scores, and accuracy at intervals. The results of the internal validation, including AUROC, Brier scores, and accuracy, can be found in Supplementary Table S4 and Supplementary Figure S1. A subgroup sensitivity analysis was performed specifically on the HBV subgroup to further examine the performance of the predictive models.

3. Results

3.1. Patients’ Baseline Characteristics

After applying the inclusion and exclusion criteria, a total of 327 patients diagnosed with the first episode of SBP were enrolled in the study. Table 1 presents the demographic characteristics of the patients, etiology of cirrhosis, hemogram and biochemical laboratory data measured at diagnosis, as well as the scores of the CTP and four MELD-based models. The patients were predominantly middle-aged and predominantly male. Among the patients, 69.2% had chronic hepatitis virus infections, with HBV infection being the most common etiology. Among the HBV-infected patients, 57.4% (85/148) received antiviral treatment. Sepsis was diagnosed in 12.8% of the SBP patients based on positive blood culture and evidence of systemic inflammatory response syndrome. Hepatorenal syndrome was present in 9.5% of the patients, and 25.7% had HE grade ≥ 1.
Table 1 also displays the interpretation of various abnormal laboratory data in patients with SBP. These abnormalities include renal function impairment, hyperbilirubinemia, hyponatremia, hypoalbuminemia, prolonged INR for prothrombin time, anemia, and thrombocytopenia. The mean white blood cell counts (WBCs) in blood were within the normal range, while the ascites polymorphonuclear leukocytes (PMNs) counts were >250 cells/μL. The mean CTP class was C. The scores for the four MELD-based models were displayed in Table 1.

3.2. Comparison of Score Differences among CTP and Four MELD-Based Models between In-Hospital, 3-Month, and 6-Month Mortality and Non-Mortality Groups

Table 2 presents the statistical differences between the in-hospital, 3-month, and 6-month mortality groups and the non-mortality group in terms of CTP and four MELD-based model scores. The in-hospital mortality rate was 39.4%. The primary causes of in-hospital mortality were sepsis with multiple organ failures (81 patients, 62.8%), end-stage liver disease (ESLD)-related multiple organ failures (28 patients, 21.7%), massive gastrointestinal bleeding with shock (13 patients, 10.1%), and other causes (7 patients, 5.4%).
The cumulative 3-month and 6-month mortality rates were 51.4% and 60.9%, respectively. Significant differences were observed between the in-hospital, 3-month, and 6-month mortality and non-mortality groups in terms of the CTP score and the four MELD-based model scores (Table 2).

3.3. Comparison of the Predicting Performance of the Original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the New SBP Scores for In-Hospital, 3-Month, and 6-Month Mortality Prediction Using AUROC and DeLong Test

To compare the predictive ability of the CTP score and the four MELD-based model scores for short-term mortality prediction (in-hospital, 3-month, and 6-month mortality), the AUROC was calculated for each score and compared.
In terms of predicting in-hospital mortality, as shown in Table 3 and Figure 2, the MELD 3.0 score exhibited the highest AUROC (0.786), followed by MELD-Na (0.783), iMELD (0.780), MELD (0.775), and CTP score (0.701). The MELD 3.0 score was significantly superior to the CTP score and non-significantly superior to MELD, MELD-Na, and iMELD scores in predicting in-hospital mortality.
In terms of predicting 3-month mortality, Table 4 and Figure 3 demonstrate that the MELD 3.0 score exhibited the highest AUROC (0.760), followed by MELD-Na (0.755), iMELD (0.753), MELD (0.737), and CTP score (0.656). The MELD 3.0 score was significantly superior to the CTP score and MELD score, and non-significantly superior to MELD-Na and iMELD scores in predicting 3-month mortality.
In terms of predicting 6-month mortality, Table 5 and Figure 4 demonstrate that the iMELD score exhibited the highest AUROC (0.752), followed by MELD-Na (0.745), MELD 3.0 (0.742), MELD (0.728), and CTP score (0.640). The iMELD score was significantly superior to the CTP score and non-significantly superior to MELD-Na, MELD 3.0, and MELD scores in predicting 6-month mortality.

3.4. Development of the New SBP Score Model

We developed a new SBP score model to enhance the accuracy of predicting in-hospital, 3-month, and 6-month mortalities. Based on the methods outlined in the Section 2, the new SBP score model was constructed by integrating age, serum creatinine, total bilirubin, International Normalized Ratio (INR), platelet count, as well as the presence of sepsis and hepatorenal syndrome (HRS).
The formula of the new SBP score model is as follows:
For in-hospital mortality: Score = −5.982 + 0.040 × Age + 0.202 × serum Cr + 0.117 × serum total bilirubin + 0.777 × INR + 0.004 × PLT + 1.009 × sepsis (0: without sepsis/1: with sepsis) + 1.576 × HRS (0: without HRS/1: with HRS).
For 3-month mortality: Score = −4.816 + 0.037 × Age + 0.132 × serum Cr + 0.136 × serum total bilirubin + 0.482 × INR + 0.005 × PLT + 0.978 × sepsis (0: without sepsis/1: with sepsis) + 1.542 × HRS (0: without HRS/1: with HRS).
For 6-month mortality: Score = −4.837 + 0.036 × Age + 0.604 × serum Cr + 0.112 × serum total bilirubin + 0.500 × INR + 0.005 × PLT + 0.661 × sepsis (0: without sepsis/1: with sepsis) + 0.858 × HRS (0: without HRS/1: with HRS).

3.5. Comparative Analysis of the Predictive Performance between the New SBP Score and the Original CTP, MELD, MELD-Na, iMELD, and MELD 3.0 Scores for In-Hospital, 3-Month, and 6-Month Mortality Prediction Utilizing AUROC and DeLong Test

According to the results presented in Table 3 and Figure 2, the newly developed SBP score model exhibited an impressive AUROC of 0.827. Notably, it demonstrated significant superiority over the CTP score, MELD, MELD-Na, and iMELD scores in predicting in-hospital mortality. Furthermore, the new SBP score model displayed a marginal advantage compared to the MELD 3.0 score in its predictive capability for in-hospital mortality.
Moreover, in terms of predicting 3-month mortality, the newly developed SBP score model demonstrated significant superiority over the CTP score and MELD (Table 4 and Figure 3). However, it did not show significant superiority over the MELD-Na, iMELD, and MELD 3.0 scores.
Additionally, when considering the prediction of 6-month mortality, the newly developed SBP score model displayed significant superiority over the CTP score (Table 5 and Figure 4). However, it did not show significant superiority over the MELD, MELD-Na, iMELD, and MELD 3.0 scores and its AUROC did not surpass 0.8.

3.6. Model Validation

The model was internally validated using five-fold cross-validation. The performance of the new model was assessed using various metrics, including AUROC, Brier scores, and accuracy. The results of the model validation, specifically for in-hospital mortality, are presented in Supplementary Table S1 and Supplementary Figure S1. The model validation results showed good performance in terms of AUROC (>0.8), Brier scores, and accuracy for the primary outcome at different time points (3- and 6-month results not shown for brevity).

3.7. Comparative Subgroup Sensitivity Analysis of the Predictive Performance between the New SBP Score and the Original CTP, MELD, MELD-Na, iMELD, and MELD 3.0 Scores for 3-Month Mortality Prediction in Patients with HBV-Related Cirrhosis and SBP

Given the high prevalence of HBV-related cirrhosis in Taiwan, a subgroup sensitivity analysis was conducted to assess the predictive accuracy of the scores in this specific patient population with SBP. In terms of 3-month mortality prediction, the new SBP score showed significant superiority over the CTP score, MELD, and iMELD scores, as indicated in Table 6 and Figure 5. However, it did not demonstrate significant superiority over the MELD-Na and MELD 3.0 scores.

3.8. The Cut-Off Values for the New SBP Score Model

The new SBP model demonstrated a cut-off value of −0.4697 for predicting in-hospital mortality, with mortality rates of 18.7% and 73.3% below and above this threshold, respectively (Supplementary Table S2). For 3-month mortality prediction, the new model’s cut-off value was 0.3712, with mortality rates of 33.6% and 84.9% below and above this value, respectively (Supplementary Table S3). Similarly, for 6-month mortality prediction, the cut-off value was 0.4690, corresponding to mortality rates of 40.8% and 83.6% below and above this threshold, respectively (Supplementary Table S4).

4. Discussion

SBP is a common and severe complication in cirrhotic patients with ascites [38], with a prevalence ranging from 10% to 30% in hospitalized patients [4,39]. Despite improvements in early recognition and effective antibiotic treatment, in-hospital mortality rates remain high, ranging from 20% to 40% [40,41]. Early identification of high-risk patients could potentially improve survival outcomes [42]. However, specific prediction models for SBP are currently unavailable. Therefore, our study aimed to compare the predictive ability of commonly used cirrhotic prediction models, including the CTP score and four MELD-based scores (MELD, MELD-Na, iMELD, MELD 3.0), for short-term mortality prediction (in-hospital, 3 and 6 month). Additionally, we sought to develop a novel model to enhance mortality prediction compared to conventional models. In this retrospective cohort study, we analyzed data from 327 cirrhotic patients who experienced their first SBP episode between January 2006 and August 2017. Clinical data were collected at the time of diagnosis, and patients were followed until March 2018 or death. Mortality prediction was assessed using AUROC analysis. HBV infection accounted for 45.3% of cirrhosis etiologies. In-hospital mortality was 39.4%, with cumulative 3-month and 6-month mortality rates of 51.4% and 60.9%, respectively. Among the existing prediction models, MELD 3.0 showed the highest AUROC in predicting in-hospital and 3-month mortality, surpassing iMELD, MELD, CTP, and MELD-Na. For 6-month mortality, iMELD showed the best performance. Our novel SBP model, incorporating age, serum Cr, total bilirubin, INR, PLT, sepsis, and HRS, demonstrated superior predictive accuracy for in-hospital mortality in all patients (AUROC 0.827) and 3-month mortality in HBV cirrhosis (AUROC 0.813), with HRS exhibiting the highest odds ratio. Internal validation and subgroup sensitivity analysis, explicitly focusing on HBV-related cirrhosis, further corroborated these findings. Our results emphasize the importance of developing a dedicated prognostic model for SBP and highlight the significance of preventing HRS development to improve the prognosis of SBP.
The CTP score has been extensively utilized for prognostic evaluation in cirrhotic patients for over four decades [43]. However, it possesses certain limitations, including the subjectivity of certain clinical parameters and its limited discriminative ability [44,45,46]. On the other hand, MELD has demonstrated superior accuracy in predicting 3-month survival compared to the CTP scoring system for cirrhotic patients awaiting liver transplantation in the United States [9]. Since February 2002, MELD has served as the primary reference system for organ allocation in liver transplantation in the United States [47]. Furthermore, MELD has shown effectiveness in predicting mortality across various chronic liver diseases and certain complications of portal hypertension, such as hepatic encephalopathy and acute variceal bleeding [48,49].
Several MELD-based models have been developed to enhance the accuracy of predicting wait-list mortality for liver transplantation. One such model is MELD-Na, which incorporates serum sodium (Na) as an additional predictor of wait-list mortality [50,51]. MELD-Na has been shown to accurately predict 6-month mortality in cirrhotic patients awaiting liver transplantation in a multicenter study [11]. Another model, the iMELD score, was developed to improve survival prediction in cirrhotic patients [12]. These MELD-based models have been compared and exhibited strong prognostic capabilities for outcome prediction in cirrhotic patients [16,17]. Recently, MELD has been updated to version 3.0 with the aim of reducing deaths while patients are on the liver transplantation waitlist [15]. However, there is limited information on the applicability of these models in the subgroup of patients with liver cirrhosis-related complications, particularly SBP, in regions with a high prevalence of chronic hepatitis B, such as Taiwan [22,52].
In our study, the MELD 3.0 score demonstrated significant superiority over the CTP score and non-significant superiority over MELD, MELD-Na, and iMELD scores in predicting in-hospital mortality. However, its AUROC of 0.786 did not reach the desired threshold of 0.8. To overcome this limitation, we developed a new SBP score model with an impressive AUROC of 0.827, surpassing the critical threshold of 0.8. The new SBP score model significantly outperformed the CTP score, MELD, MELD-Na, and iMELD scores, and slightly outperformed the MELD 3.0 score in predicting in-hospital mortality. For the 3-month mortality prediction, the new SBP score model demonstrated significant superiority over the CTP score and MELD, with an AUROC of 0.789. Subgroup sensitivity analysis focusing on HBV-related cirrhosis and SBP further improved the AUROC to 0.813. However, the performance of the new SBP score model in predicting 6-month mortality was not satisfactory, with an AUROC of 0.762, similar to that of the iMELD score. These findings suggest that the new SBP score model is particularly effective in predicting in-hospital and 3-month mortality, aligning with the current policy of assessing the risk of death within 90 days for liver transplants [53].
It is noteworthy that our study’s findings align with previous research emphasizing the detrimental impact of HRS on SBP prognosis. This is supported by the highest odds ratio in our SBP score model, highlighting the significant role of HRS in predicting SBP outcomes. These findings are consistent with a well-known study demonstrating the benefits of intravenous albumin treatment in conjunction with antibiotics, which reduces the incidence of renal impairment and mortality in cirrhotic patients with SBP, compared to antibiotic treatment alone [54]. This finding underscores the importance of preventing HRS development and highlights the potential value of interventions targeting renal function in improving SBP prognosis.
The study has several limitations. Firstly, it is a retrospective study with a relatively small number of patients. Secondly, external validation was not conducted. Thirdly, being a single-center study, a multi-center study would be beneficial in the future. Despite these limitations, the findings underscore the need for a specific prognostic model for SBP and emphasize the importance of conducting a prospective and larger case series for validation.

5. Conclusions

In our study of cirrhotic patients with the first episode of SBP, the MELD 3.0 score outperformed the CTP score and showed a non-significant improvement compared to other MELD-based scores (MELD, MELD-Na, iMELD) in predicting in-hospital and 3-month mortality, although the AUROC values were below the desired threshold of 0.8. Our newly developed SBP model, incorporating age, serum Cr, total bilirubin, INR, PLT, sepsis, and HRS, achieved an impressive AUROC value exceeding 0.8 for in-hospital mortality in all patients and 3-month mortality in HBV cirrhosis. Internal validation and subgroup sensitivity analysis focusing on HBV-related cirrhosis supported these findings. These results highlight the need for a specific prognostic model for SBP and underscored the importance of preventing HRS development to improve SBP prognosis.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/diagnostics13152578/s1. Table S1: Model internal validation by 5-fold cross-validation. Table S2: The cut-off values for the new SBP score model in predicting in-hospital mortality. Table S3: The cut-off values for the new SBP score model in predicting 3-month mortality. Table S4: The cut-off values for the new SBP score model in predicting 6-month mortality. Figure S1: The AUROC curve of the training set and the test (validation) set for predicting in-hospital mortality.

Author Contributions

Formal analysis, Y.L. and L.-T.L.; Investigation, W.-T.C., W.T., Y.-C.H. and Y.-M.W.; Resources, T.-H.W.; Writing—original draft, Y.-T.L.; Writing—review & editing, C.-H.H.; Supervision, C.-W.H. and R.-N.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Chang Gung Medical Foundation grant number CMRPG3M1931-32.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Chang Gung Memorial Hospital (202000112B0; date of approval: 5 February 2020).

Informed Consent Statement

Patient consent was waived by the IRB as this is a retrospective cohort study. The study involved only medical chart reviews without any patient intervention or the disclosure of personal information. Hence, the IRB did not require patients’ signatures.

Data Availability Statement

Data and study materials will be made available to other researchers upon request via email, along with their IRB approval document.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

ALTalanine aminotransferase
ASTaspartate aminotransferase
AUROCarea under the receiver operating characteristic curve
GIgastrointestinal
HCChepatocellular carcinoma
HCVhepatitis C virus
HRShepatorenal syndrome
LCliver cirrhosis
PLTplatelet
PMNpolymorphonuclear neutrophils
PPVpositive predictive value
PT-INRProthrombin time–International Normalized Ratio
SBPspontaneous bacterial peritonitis
WBCwhite blood cell

References

  1. Wiest, R.; Lawson, M.; Geuking, M. Pathological bacterial translocation in liver cirrhosis. J. Hepatol. 2014, 60, 197–209. [Google Scholar] [CrossRef] [Green Version]
  2. Fernandez, J.; Navasa, M.; Gomez, J.; Colmenero, J.; Vila, J.; Arroyo, V.; Rodes, J. Bacterial infections in cirrhosis: Epidemiological changes with invasive procedures and norfloxacin prophylaxis. Hepatology 2002, 35, 140–148. [Google Scholar] [CrossRef]
  3. Gines, P.; Cardenas, A.; Arroyo, V.; Rodes, J. Management of cirrhosis and ascites. N. Engl. J. Med. 2004, 350, 1646–1654. [Google Scholar] [CrossRef] [Green Version]
  4. Huang, C.; Lee, C.; Chang, C. Spontaneous Bacterial Peritonitis in Decompensated Liver Cirrhosis—A Literature Review. Livers 2022, 2, 214–232. [Google Scholar] [CrossRef]
  5. Biggins, S.W.; Angeli, P.; Garcia-Tsao, G.; Gines, P.; Ling, S.C.; Nadim, M.K.; Wong, F.; Kim, W.R. Diagnosis, Evaluation, and Management of Ascites, Spontaneous Bacterial Peritonitis and Hepatorenal Syndrome: 2021 Practice Guidance by the American Association for the Study of Liver Diseases. Hepatology 2021, 74, 1014–1048. [Google Scholar] [CrossRef] [PubMed]
  6. Hung, T.H.; Tsai, C.C.; Hsieh, Y.H.; Tsai, C.C. The long-term mortality of spontaneous bacterial peritonitis in cirrhotic patients: A 3-year nationwide cohort study. Turk. J. Gastroenterol. 2015, 26, 159–162. [Google Scholar] [CrossRef]
  7. Andreu, M.; Sola, R.; Sitges-Serra, A.; Alia, C.; Gallen, M.; Vila, M.C.; Coll, S.; Oliver, M.I. Risk factors for spontaneous bacterial peritonitis in cirrhotic patients with ascites. Gastroenterology 1993, 104, 1133–1138. [Google Scholar] [CrossRef] [PubMed]
  8. Huang, C.H.; Tseng, H.J.; Amodio, P.; Chen, Y.L.; Wang, S.F.; Chang, S.H.; Hsieh, S.Y.; Lin, C.Y. Hepatic Encephalopathy and Spontaneous Bacterial Peritonitis Improve Cirrhosis Outcome Prediction: A Modified Seven-Stage Model as a Clinical Alternative to MELD. J. Pers. Med. 2020, 10, 186. [Google Scholar] [CrossRef]
  9. Wiesner, R.; Edwards, E.; Freeman, R.; Harper, A.; Kim, R.; Kamath, P.; Kremers, W.; Lake, J.; Howard, T.; Merion, R.M.; et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003, 124, 91–96. [Google Scholar] [CrossRef] [Green Version]
  10. Asrani, S.K.; Kim, W.R. Model for end-stage liver disease: End of the first decade. Clin. Liver Dis. 2011, 15, 685–698. [Google Scholar] [CrossRef] [Green Version]
  11. Biggins, S.W.; Kim, W.R.; Terrault, N.A.; Saab, S.; Balan, V.; Schiano, T.; Benson, J.; Therneau, T.; Kremers, W.; Wiesner, R.; et al. Evidence-based incorporation of serum sodium concentration into MELD. Gastroenterology 2006, 130, 1652–1660. [Google Scholar] [CrossRef]
  12. Luca, A.; Angermayr, B.; Bertolini, G.; Koenig, F.; Vizzini, G.; Ploner, M.; Peck-Radosavljevic, M.; Gridelli, B.; Bosch, J. An integrated MELD model including serum sodium and age improves the prediction of early mortality in patients with cirrhosis. Liver Transplant. 2007, 13, 1174–1180. [Google Scholar] [CrossRef]
  13. Huo, T.I.; Wang, Y.W.; Yang, Y.Y.; Lin, H.C.; Lee, P.C.; Hou, M.C.; Lee, F.Y.; Lee, S.D. Model for end-stage liver disease score to serum sodium ratio index as a prognostic predictor and its correlation with portal pressure in patients with liver cirrhosis. Liver Int. 2007, 27, 498–506. [Google Scholar]
  14. Leise, M.D.; Kim, W.R.; Kremers, W.K.; Larson, J.J.; Benson, J.T.; Therneau, T.M. A revised model for end-stage liver disease optimizes prediction of mortality among patients awaiting liver transplantation. Gastroenterology 2011, 140, 1952–1960. [Google Scholar]
  15. Kim, W.R.; Mannalithara, A.; Heimbach, J.K.; Kamath, P.S.; Asrani, S.K.; Biggins, S.W.; Wood, N.L.; Gentry, S.E.; Kwong, A.J. MELD 3.0: The Model for End-Stage Liver Disease Updated for the Modern Era. Gastroenterology 2021, 161, 1887–1895.E4. [Google Scholar] [CrossRef] [PubMed]
  16. Huo, T.I.; Lin, H.C.; Huo, S.C.; Lee, P.C.; Wu, J.C.; Lee, F.Y.; Hou, M.C.; Lee, S.D. Comparison of four model for end-stage liver disease-based prognostic systems for cirrhosis. Liver Transplant. 2008, 14, 837–844. [Google Scholar] [CrossRef]
  17. Koo, J.K.; Kim, J.H.; Choi, Y.J.; Lee, C.I.; Yang, J.H.; Yoon, H.Y.; Choi, H.J.; Ko, S.Y.; Choe, W.H.; Kwon, S.Y.; et al. Predictive value of Refit Model for End-Stage Liver Disease, Refit Model for End-Stage Liver Disease-Na, and pre-existing scoring system for 3-month mortality in Korean patients with cirrhosis. J. Gastroenterol. Hepatol. 2013, 28, 1209–1216. [Google Scholar] [CrossRef] [PubMed]
  18. Huo, T.I.; Lin, H.C.; Wu, J.C.; Hou, M.C.; Lee, F.Y.; Lee, P.C.; Chang, F.Y.; Lee, S.D. Limitation of the model for end-stage liver disease for outcome prediction in patients with cirrhosis-related complications. Clin. Transplant. 2006, 20, 188–194. [Google Scholar] [CrossRef] [PubMed]
  19. Chawla, Y.K.; Kashinath, R.C.; Duseja, A.; Dhiman, R.K. Predicting Mortality Across a Broad Spectrum of Liver Disease—An Assessment of Model for End-Stage Liver Disease (MELD), Child-Turcotte-Pugh (CTP), and Creatinine-Modified CTP Scores. J. Clin. Exp. Hepatol. 2011, 1, 161–168. [Google Scholar] [CrossRef] [Green Version]
  20. Liaw, Y.F.; Chu, C.M. Hepatitis B virus infection. Lancet 2009, 373, 582–592. [Google Scholar] [CrossRef]
  21. Huang, C.H.; Hsieh, S.Y. More immunosuppressive, more immunotherapy responsive? A double-edged sword of HBV-induced immune response in HCC. Hepatology, 2023; online ahead of print. [Google Scholar] [CrossRef]
  22. Chen, P.C.; Chen, B.H.; Huang, C.H.; Jeng, W.J.; Hsieh, Y.C.; Teng, W.; Chen, Y.C.; Ho, Y.P.; Sheen, I.S.; Lin, C.Y. Integrated model for end-stage liver disease maybe superior to some other model for end-stage liver disease-based systems in addition to Child-Turcotte-Pugh and albumin-bilirubin scores in patients with hepatitis B virus-related liver cirrhosis and spontaneous bacterial peritonitis. Eur. J. Gastroenterol. Hepatol. 2019, 31, 1256–1263. [Google Scholar] [CrossRef] [PubMed]
  23. Huang, C.H.; Wang, S.F.; Lee, C.H.; Wu, Y.M.; Chang, C.; Chen, B.H.; Huang, Y.T.; Ho, Y.P. Bacteremia (Sepsis), Hepatorenal Syndrome, and Serum Creatinine Levels Rather than Types or Microbial Patterns Predicted the Short-Term Survival of Cirrhotic Patients Complicated with Spontaneous Bacterial Peritonitis. Diagnostics 2023, 13, 94. [Google Scholar] [CrossRef]
  24. Lin, D.Y.; Sheen, I.S.; Chiu, C.T.; Lin, S.M.; Kuo, Y.C.; Liaw, Y.F. Ultrasonographic changes of early liver cirrhosis in chronic hepatitis B: A longitudinal study. J. Clin. Ultrasound 1993, 21, 303–308. [Google Scholar] [CrossRef] [PubMed]
  25. Tapper, E.B.; Parikh, N.D. Diagnosis and Management of Cirrhosis and Its Complications: A Review. JAMA 2023, 329, 1589–1602. [Google Scholar] [CrossRef] [PubMed]
  26. Runyon, B.A.; Montano, A.A.; Akriviadis, E.A.; Antillon, M.R.; Irving, M.A.; McHutchison, J.G. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann. Intern. Med. 1992, 117, 215–220. [Google Scholar] [CrossRef]
  27. Runyon, B.A. Management of adult patients with ascites caused by cirrhosis. Hepatology 1998, 27, 264–272. [Google Scholar] [CrossRef]
  28. Rimola, A.; García-Tsao, G.; Navasa, M.; Piddock, L.J.; Planas, R.; Bernard, B.; Inadomi, J.M. Diagnosis, treatment and prophylaxis of spontaneous bacterial peritonitis: A consensus document. J. Hepatol. 2000, 32, 142–153. [Google Scholar]
  29. Lee, Y.Y.; Tee, H.P.; Mahadeva, S. Role of prophylactic antibiotics in cirrhotic patients with variceal bleeding. World J. Gastroenterol. 2014, 20, 1790–1796. [Google Scholar] [CrossRef]
  30. European Association for the Study of the Liver. EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis. J. Hepatol. 2018, 69, 406–460. [Google Scholar] [CrossRef] [Green Version]
  31. Angeli, P.; Gines, P.; Wong, F.; Bernardi, M.; Boyer, T.D.; Gerbes, A.; Moreau, R.; Jalan, R.; Sarin, S.K.; Piano, S.; et al. Diagnosis and management of acute kidney injury in patients with cirrhosis: Revised consensus recommendations of the International Club of Ascites. J. Hepatol. 2015, 62, 968–974. [Google Scholar] [CrossRef] [Green Version]
  32. Liaw, Y.F.; Leung, N.; Kao, J.H.; Piratvisuth, T.; Gane, E.; Han, K.H.; Guan, R.; Lau, G.K.; Locarnini, S. Asian-Pacific consensus statement on the management of chronic hepatitis B: A 2008 update. Hepatol. Int. 2008, 2, 263–283. [Google Scholar] [CrossRef] [Green Version]
  33. Liaw, Y.F.; Kao, J.H.; Piratvisuth, T.; Chan, H.L.; Chien, R.N.; Liu, C.J.; Gane, E.; Locarnini, S.; Lim, S.G.; Han, K.H.; et al. Asian-Pacific consensus statement on the management of chronic hepatitis B: A 2012 update. Hepatol. Int. 2012, 6, 531–561. [Google Scholar] [CrossRef] [Green Version]
  34. Sarin, S.K.; Kumar, M.; Lau, G.K.; Abbas, Z.; Chan, H.L.; Chen, C.J.; Chen, D.S.; Chen, H.L.; Chen, P.J.; Chien, R.N.; et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: A 2015 update. Hepatol. Int. 2016, 10, 1–98. [Google Scholar] [CrossRef]
  35. Pugh, R.; Murray-Lyon, I.; Dawson, J.; Pietroni, M.; Williams, R. Transection of the oesophagus for bleeding oesophageal varices. Br. J. Surg. 1973, 60, 646–649. [Google Scholar]
  36. Maldonado, G.; Greenland, S. Simulation study of confounder-selection strategies. Am. J. Epidemiol. 1993, 138, 923–936. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  37. Sarmast, N.; Ogola, G.O.; Kouznetsova, M.; Leise, M.D.; Bahirwani, R.; Maiwall, R.; Tapper, E.; Trotter, J.; Bajaj, J.S.; Thacker, L.R.; et al. Model for End-Stage Liver Disease-Lactate and Prediction of Inpatient Mortality in Patients With Chronic Liver Disease. Hepatology 2020, 72, 1747–1757. [Google Scholar] [CrossRef]
  38. Lata, J.; Stiburek, O.; Kopacova, M. Spontaneous bacterial peritonitis: A severe complication of liver cirrhosis. World J. Gastroenterol. 2009, 15, 5505–5510. [Google Scholar] [CrossRef]
  39. Aithal, G.P.; Palaniyappan, N.; China, L.; Harmala, S.; Macken, L.; Ryan, J.M.; Wilkes, E.A.; Moore, K.; Leithead, J.A.; Hayes, P.C.; et al. Guidelines on the management of ascites in cirrhosis. Gut 2021, 70, 9–29. [Google Scholar] [CrossRef] [PubMed]
  40. Thuluvath, P.J.; Morss, S.; Thompson, R. Spontaneous bacterial peritonitis—In-hospital mortality, predictors of survival, and health care costs from 1988 to 1998. Am. J. Gastroenterol. 2001, 96, 1232–1236. [Google Scholar] [CrossRef]
  41. Huang, C.H.; Lin, C.Y.; Sheen, I.S.; Chen, W.T.; Lin, T.N.; Ho, Y.P.; Chiu, C.T. Recurrence of spontaneous bacterial peritonitis in cirrhotic patients non-prophylactically treated with norfloxacin: Serum albumin as an easy but reliable predictive factor. Liver Int. 2011, 31, 184–191. [Google Scholar] [CrossRef] [PubMed]
  42. Garcia-Tsao, G. Current management of the complications of cirrhosis and portal hypertension: Variceal hemorrhage, ascites, and spontaneous bacterial peritonitis. Gastroenterology 2001, 120, 726–748. [Google Scholar] [CrossRef]
  43. Kaplan, D.E.; Dai, F.; Aytaman, A.; Baytarian, M.; Fox, R.; Hunt, K.; Knott, A.; Pedrosa, M.; Pocha, C.; Mehta, R.; et al. Development and Performance of an Algorithm to Estimate the Child-Turcotte-Pugh Score From a National Electronic Healthcare Database. Clin. Gastroenterol. Hepatol. 2015, 13, 2333–2341.E6. [Google Scholar] [CrossRef] [Green Version]
  44. Oellerich, M.; Burdelski, M.; Lautz, H.U.; Rodeck, B.; Duewel, J.; Schulz, M.; Schmidt, F.W.; Brodehl, J.; Pichlmayr, R. Assessment of pretransplant prognosis in patients with cirrhosis. Transplantation 1991, 51, 801–806. [Google Scholar] [CrossRef] [PubMed]
  45. Shrestha, R.; McKinley, C.; Showalter, R.; Wilner, K.; Marsano, L.; Vivian, B.; Everson, G.T. Quantitative liver function tests define the functional severity of liver disease in early-stage cirrhosis. Liver Transplant. Surg. 1997, 3, 166–173. [Google Scholar] [CrossRef]
  46. Testa, R.; Valente, U.; Risso, D.; Caglieris, S.; Giannini, E.; Fasoli, A.; Botta, F.; Dardano, G.; Lantieri, P.B.; Celle, G. Can the MEGX test and serum bile acids improve the prognostic ability of Child-Pugh’s score in liver cirrhosis? Eur. J. Gastroenterol. Hepatol. 1999, 11, 559–563. [Google Scholar] [CrossRef] [PubMed]
  47. Olthoff, K.M.; Brown, R.S., Jr.; Delmonico, F.L.; Freeman, R.B.; McDiarmid, S.V.; Merion, R.M.; Millis, J.M.; Roberts, J.P.; Shaked, A.; Wiesner, R.H.; et al. Summary report of a national conference: Evolving concepts in liver allocation in the MELD and PELD era. Liver Transplant. 2004, 10 (Suppl. S2), A6–A22. [Google Scholar] [CrossRef]
  48. Said, A.; Williams, J.; Holden, J.; Remington, P.; Gangnon, R.; Musat, A.; Lucey, M.R. Model for end stage liver disease score predicts mortality across a broad spectrum of liver disease. J. Hepatol. 2004, 40, 897–903. [Google Scholar] [CrossRef]
  49. Chalasani, N.; Kahi, C.; Francois, F.; Pinto, A.; Marathe, A.; Bini, E.J.; Pandya, P.; Sitaraman, S.; Shen, J. Model for end-stage liver disease (MELD) for predicting mortality in patients with acute variceal bleeding. Hepatology 2002, 35, 1282–1284. [Google Scholar] [CrossRef]
  50. Biggins, S.W.; Rodriguez, H.J.; Bacchetti, P.; Bass, N.M.; Roberts, J.P.; Terrault, N.A. Serum sodium predicts mortality in patients listed for liver transplantation. Hepatology 2005, 41, 32–39. [Google Scholar] [CrossRef]
  51. Ruf, A.E.; Kremers, W.K.; Chavez, L.L.; Descalzi, V.I.; Podesta, L.G.; Villamil, F.G. Addition of serum sodium into the MELD score predicts waiting list mortality better than MELD alone. Liver Transplant. 2005, 11, 336–343. [Google Scholar] [CrossRef]
  52. Yang, H.I.; Lu, S.N.; Liaw, Y.F.; You, S.L.; Sun, C.A.; Wang, L.Y.; Hsiao, C.K.; Chen, P.J.; Chen, D.S.; Chen, C.J.; et al. Hepatitis B e antigen and the risk of hepatocellular carcinoma. N. Engl. J. Med. 2002, 347, 168–174. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  53. Kwong, A.J.; Zhang, K.Y.; Ebel, N.; Mannalithara, A.; Kim, W.R. MELD 3.0 for adolescent liver transplant candidates. Hepatology 2023, 78, 540–546. [Google Scholar] [CrossRef] [PubMed]
  54. Sort, P.; Navasa, M.; Arroyo, V.; Aldeguer, X.; Planas, R.; Ruiz-del-Arbol, L.; Castells, L.; Vargas, V.; Soriano, G.; Guevara, M.; et al. Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis. N. Engl. J. Med. 1999, 341, 403–409. [Google Scholar] [CrossRef] [PubMed] [Green Version]
Diagnostics 13 02578 g001
Figure 1. Flowchart of the study. Following the inclusion and exclusion criteria, a total of 327 patients diagnosed with the first episode of SBP were included in the study analysis.
Figure 1. Flowchart of the study. Following the inclusion and exclusion criteria, a total of 327 patients diagnosed with the first episode of SBP were included in the study analysis.
Diagnostics 13 02578 g001
Diagnostics 13 02578 g002
Figure 2. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting in-hospital mortality based on AUROC. The MELD 3.0 score showed significant superiority over the CTP score and non-significant improvement over MELD, MELD-Na, and iMELD scores. The newly developed SBP score exhibited an impressive AUROC of 0.827 and demonstrated significant superiority over the CTP score, MELD, MELD-Na, and iMELD scores. Furthermore, the new SBP score displayed a marginal advantage compared to the MELD 3.0 score.
Figure 2. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting in-hospital mortality based on AUROC. The MELD 3.0 score showed significant superiority over the CTP score and non-significant improvement over MELD, MELD-Na, and iMELD scores. The newly developed SBP score exhibited an impressive AUROC of 0.827 and demonstrated significant superiority over the CTP score, MELD, MELD-Na, and iMELD scores. Furthermore, the new SBP score displayed a marginal advantage compared to the MELD 3.0 score.
Diagnostics 13 02578 g002
Diagnostics 13 02578 g003
Figure 3. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality using AUROC. The MELD 3.0 score exhibited the highest AUROC, followed by MELD-Na, iMELD, MELD, and CTP scores. The MELD 3.0 score was significantly superior to the CTP score and MELD score, and non-significantly superior to MELD-Na and iMELD scores. The newly developed SBP score demonstrated significant superiority over the CTP score and MELD. However, it did not show significant superiority over the MELD-Na, iMELD, and MELD 3.0 scores.
Figure 3. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality using AUROC. The MELD 3.0 score exhibited the highest AUROC, followed by MELD-Na, iMELD, MELD, and CTP scores. The MELD 3.0 score was significantly superior to the CTP score and MELD score, and non-significantly superior to MELD-Na and iMELD scores. The newly developed SBP score demonstrated significant superiority over the CTP score and MELD. However, it did not show significant superiority over the MELD-Na, iMELD, and MELD 3.0 scores.
Diagnostics 13 02578 g003
Diagnostics 13 02578 g004
Figure 4. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 6-month mortality by AUROC. The iMELD score exhibited the highest AUROC, followed by MELD-Na, MELD 3.0, MELD, and CTP scores. The iMELD score was significantly superior to the CTP score and non-significantly superior to MELD-Na, MELD 3.0, and MELD scores. The newly developed SBP score displayed significant superiority over the CTP score. However, it did not show significant superiority over the MELD, MELD-Na, iMELD, and MELD 3.0 scores.
Figure 4. Comparison of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 6-month mortality by AUROC. The iMELD score exhibited the highest AUROC, followed by MELD-Na, MELD 3.0, MELD, and CTP scores. The iMELD score was significantly superior to the CTP score and non-significantly superior to MELD-Na, MELD 3.0, and MELD scores. The newly developed SBP score displayed significant superiority over the CTP score. However, it did not show significant superiority over the MELD, MELD-Na, iMELD, and MELD 3.0 scores.
Diagnostics 13 02578 g004
Diagnostics 13 02578 g005
Figure 5. Comparison of the predictive performance of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality in the HBV subgroup using AUROC. The new SBP score demonstrated significant superiority (0.813) over the CTP score, MELD, and iMELD scores. However, it did not show significant superiority over the MELD-Na and MELD 3.0 scores.
Figure 5. Comparison of the predictive performance of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality in the HBV subgroup using AUROC. The new SBP score demonstrated significant superiority (0.813) over the CTP score, MELD, and iMELD scores. However, it did not show significant superiority over the MELD-Na and MELD 3.0 scores.
Diagnostics 13 02578 g005
Table 1. Demographic and baseline clinical characteristics of 327 patients with the first episode of SBP.
Table 1. Demographic and baseline clinical characteristics of 327 patients with the first episode of SBP.
Baseline ParametersValues
Clinical Parameters
Age, mean ± SD57.1 ± 13.6
Male No. (%)236 (72%)
Etiology No. (%)
Alcohol90 (27.5%)
HBV148 (45.3%)
HCV78 (23.9%)
Others11 (3.4%)
Blood culture positive42(12.8%)
Hepatorenal syndrome31(9.5%)
Hepatic encephalopathy
Grade 0243(74.3%)
Grade 129(8.9%)
Grade 232(9.8%)
Grade 315(4.6%)
Grade 48(2.4%)
Ascites
Mild43(13.2%)
Moderate76(23.2%)
Severe208(63.6%)
Laboratory Parameters
Median (IQR)
Creatinine (mg/dL)1.2 (0.86–2.41)
Bilirubin Total (mg/dL)4.1 (1.9–9.8)
Sodium (mEq/L)135 (131–139)
Albumin (g/dL)2.4 (2.2–2.8)
INR1.64 (1.39–2.20)
WBC (103/μL)7.9 (5.20–12.80)
Hemoglobin (g/dL)9.6 (8.4–11.0)
PLT (103/μL)74.0 (48.0–122.0)
Ascites PMN (cells/mm3)1393.5 (344.3–4203.0)
CTP score10.1 ± 1.9
MELD score22.7 ± 9.3
MELD-Na score25.1 ± 8.5
iMELD score45.6 ± 10.4
MELD 3.0 score25.4 ± 8.5
Table 2. Statistical differences in scores among CTP and four MELD-based models for predicting in-hospital, 3-month, and 6-month mortalities between mortality and non-mortality groups.
Table 2. Statistical differences in scores among CTP and four MELD-based models for predicting in-hospital, 3-month, and 6-month mortalities between mortality and non-mortality groups.
MortalityNon-Mortalityp
In-hospitaln = 129 (39.4%)n = 198 (60.6%)
CTP score10.95 ± 1.7859.58 ± 1.839<0.001
MELD score28.3977 ± 10.2856219.0414 ± 6.37866<0.001
MELD-Na score30.3214 ± 8.5961621.7216 ± 6.57153<0.001
iMELD score51.8581 ± 10.1740541.4468 ± 8.26332<0.001
MELD 3.0 score30.6357 ± 8.5054621.9747 ± 6.52013<0.001
3-Monthn = 168 (51.4%)n = 159 (48.6%)
CTP score10.64 ± 1.8659.57 ± 1.861<0.001
MELD score26.5179 ± 10.1902718.7327 ± 6.21459<0.001
MELD-Na score28.7469 ± 8.7490021.2759 ± 6.36454<0.001
iMELD score50.0182 ± 10.5008640.8371 ± 790264<0.001
MELD 3.0 score29.0714 ± 8.7016921.5031 ± 6.24892<0.001
6-Monthn = 199 (60.9%)n = 128 (39.1%)
CTP score10.50 ± 1.8649.52 ± 1.903<0.001
MELD score25.5815 ± 9.9027018.3030 ± 6.18074<0.001
MELD-Na score27.8953 ± 8.736420.7905 ± 6.43326<0.001
iMELD score49.1069 ± 10.3456740.0304 ± 7.71442<0.001
MELD 3.0 score28.1457 ± 8.5229321.1094 ± 6.43224<0.001
n = number of patients.
Table 3. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting in-hospital mortality by AUROC and DeLong test.
Table 3. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting in-hospital mortality by AUROC and DeLong test.
Predicting In-Hospital Mortality
CTP ScoreMELD ScoreMELD-Na ScoreiMELD ScoreMELD 3.0 ScoreNew SBP Score
AUROC0.7010.7750.7830.7800.7860.827
(95% CI)0.64–0.760.72–0.830.73–0.830.73–0.830.73–0.840.78–0.87
p-value<0.001<0.001<0.001<0.001<0.001<0.001
CTP score
MELD score0.0013
MELD-Na score0.00030.4608
iMELD score0.00410.78430.8778
MELD 3.0 score0.00010.27890.58140.7370
New SBP score<0.00010.01960.04710.02400.0651
Table 4. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality by AUROC and DeLong test.
Table 4. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality by AUROC and DeLong test.
Predicting 3-Month Mortality
CTP ScoreMELD ScoreMELD-Na ScoreiMELD ScoreMELD 3.0 ScoreNew SBP Score
AUROC0.6560.7370.7550.7530.7600.789
(95% CI)0.60–0.710.68–0.790.70–0.810.70–0.810.71–0.810.74–0.84
p-value<0.001<0.001<0.001<0.001<0.001<0.001
CTP score
MELD score0.0007
MELD-Na score<0.00010.0799
iMELD score0.00070.40110.8833
MELD 3.0 score<0.00010.02090.40020.6702
New SBP score<0.00010.02660.14990.10740.2160
Table 5. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 6-month mortality by AUROC and DeLong test.
Table 5. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 6-month mortality by AUROC and DeLong test.
Predicting 6-Month Mortality
CTP ScoreMELD ScoreMELD-Na ScoreiMELD ScoreMELD 3.0 ScoreNew SBP Score
AUROC0.6400.7280.7450.7520.7420.762
(95% CI)0.58–0.700.67–0.780.69–0.800.70–0.800.69–0.800.71–0.81
p-value<0.001<0.001<0.001<0.001<0.001<0.001
CTP score
MELD score0.0007
MELD-Na score<0.00010.0623
iMELD score0.00010.13450.5149
MELD 3.0 score<0.00010.07720.89660.4950
New SBP score0.00080.11150.43300.66020.4094
Table 6. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality in the HBV-related cirrhosis subgroup.
Table 6. Comparison of the predicting ability of the original CTP, MELD, MELD-Na, iMELD, MELD 3.0, and the new SBP scores in predicting 3-month mortality in the HBV-related cirrhosis subgroup.
Predicting 3-Month Mortality in HBV Subgroup
CTP ScoreMELD ScoreMELD-Na ScoreiMELD ScoreMELD 3.0 ScoreNew SBP Score
AUC0.7000.7550.7600.7360.7650.813
(95% CI)0.62–0.780.68–0.830.68–0.840.66–0.82069–0.840.73–0.89
p-value<0.001<0.001<0.001<0.001<0.001<0.001
CTP score
MELD score0.1442
MELD-Na score0.10140.6895
iMELD score0.42450.52330.2928
MELD 3.0 score0.07110.48270.55690.2254
New SBP score0.02910.04760.08860.04460.2035
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Lin, Y.-T.; Chen, W.-T.; Wu, T.-H.; Liu, Y.; Liu, L.-T.; Teng, W.; Hsieh, Y.-C.; Wu, Y.-M.; Huang, C.-H.; Hsu, C.-W.; et al. A Validated Composite Score Demonstrates Potential Superiority to MELD-Based Systems in Predicting Short-Term Survival in Patients with Liver Cirrhosis and Spontaneous Bacterial Peritonitis—A Preliminary Study. Diagnostics 2023, 13, 2578. https://doi.org/10.3390/diagnostics13152578

AMA Style

Lin Y-T, Chen W-T, Wu T-H, Liu Y, Liu L-T, Teng W, Hsieh Y-C, Wu Y-M, Huang C-H, Hsu C-W, et al. A Validated Composite Score Demonstrates Potential Superiority to MELD-Based Systems in Predicting Short-Term Survival in Patients with Liver Cirrhosis and Spontaneous Bacterial Peritonitis—A Preliminary Study. Diagnostics. 2023; 13(15):2578. https://doi.org/10.3390/diagnostics13152578

Chicago/Turabian Style

Lin, Yan-Ting, Wei-Ting Chen, Tsung-Han Wu, Yu Liu, Li-Tong Liu, Wei Teng, Yi-Chung Hsieh, Yen-Mu Wu, Chien-Hao Huang, Chao-Wei Hsu, and et al. 2023. "A Validated Composite Score Demonstrates Potential Superiority to MELD-Based Systems in Predicting Short-Term Survival in Patients with Liver Cirrhosis and Spontaneous Bacterial Peritonitis—A Preliminary Study" Diagnostics 13, no. 15: 2578. https://doi.org/10.3390/diagnostics13152578

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop