Risk prediction models for hepatic encephalopathy in patients with liver cirrhosis: A systematic review and meta-analysis

Risk prediction models for hepatic encephalopathy in patients with liver cirrhosis: A systematic review and meta-analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Risk prediction models for hepatic encephalopathy in patients with liver cirrhosis: A systematic review and meta-analysis Zhile Xiao, Xiaonan Li, Feng Zhou This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7875969/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background To conduct a systematic review and meta-analysis of existing prediction models for hepatic encephalopathy (HE) in patients with cirrhosis. Methods China National Knowledge Infrastructure (CNKI), WanFang Database, China Science and Technology Journal Database (VIP), SinoMed, PubMed, Embase, Web of Science, Scopus, and the Cochrane Library databases were searched for studies on prediction models for the risk of HE in cirrhosis from inception to April 12, 2025. Two researchers independently conducted the literature search and data extraction, and the quality of the literature was evaluated using the Prediction Model Risk of Bias Assessment Tool (PROBAST)—meta-analysis using Review Manager 5.4 and Stata 17.0 software. Results Thirty-eight best prediction models from thirty-eight studies were ultimately included in this review. Among them, 17 studies predicted HE after transjugular intrahepatic portosystemic shunt (TIPS). The incidence of HE ranged from 7.2% to 50.4%. The most commonly used predictors were age and Child-Pugh grade/score. The reported area under the curve (AUC) or c-statistic values ranged from 0.667 to 0.969. Thirty-four studies were found to have a high risk of bias, and 27 studies raised applicability concerns, primarily due to inappropriate data sources, limitations in the domain of analysis, and homogenous study populations. Four externally validated logistic regression models had a combined AUC of 0.802 (95% CI: 0.785–0.820), indicating moderate predictive performance. In meta-analysis, age (OR = 1.04, 95% CI: 1.03, 1.05), prior HE (OR = 4.42, 95% CI: 2.67, 7.31), low albumin (OR = 1.78, 95% CI: 1.25, 2.56), total bilirubin (OR = 2.22, 95% CI: 1.73, 2.85), Child-Pugh grade/score (OR = 2.41, 95% CI: 1.87, 3.09; OR = 1.65, 95% CI: 1.16, 2.33, respectively), ascites (OR = 1.96, 95% CI: 1.48, 2.60), and co-infection (OR = 2.57, 95% CI: 1.66, 3.98) were significant predictors of HE in cirrhosis (P < 0.01). Conclusions Prediction models for estimating the risk of incident HE with cirrhosis demonstrate moderate discrimination performance, while with a high overall risk of bias and a lack of clinical effectiveness research. Future research should focus on developing new models through optimised study design and analysis, increased sample sizes and external validation, and applying them to clinical practice. Registration: The protocol for this review was registered on PROSPERO (CRD420251040913). liver cirrhosis hepatic encephalopathy risk prediction model systematic review meta-analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Hepatic encephalopathy (HE) is a neuropsychiatric complication arising from various acute and chronic liver diseases or abnormalities of portal-corporeal circulation shunting[ 1 ]. In terms of its severity, the spectrum of HE ranges from subclinical changes in cognitive function, known as covert or minimal HE (CHE or MHE), to overt HE (OHE) with disorientation, confusion, and coma[ 2 , 3 ]. It is estimated that 30% − 40% of patients with cirrhosis will experience at least one episode of OHE during their lifetime, and the recurrence rate following an initial episode remains high[ 4 , 5 ]. The burdens and costs associated with HE for patients and the healthcare system are extensive and increasing[ 5 ]. Numerous risk factors contribute to the development of HE. These factors include the severity of liver dysfunction (often assessed by Child-Pugh or model for end-stage liver disease (MELD) scores), a history of prior HE episodes, the presence of a transjugular intrahepatic portosystemic shunt (TIPS), gastrointestinal bleeding, infections, electrolyte disturbances (such as hyponatremia and hypokalemia), constipation, advanced age, and the use of specific medications like sedatives or diuretics[ 1 ]. In addition, recent studies have emphasized the predictive value of factors such as sarcopenia, genetic factors, and specific laboratory indicators[ 6 – 8 ]. Various prediction models for HE in cirrhosis patients have been developed and evaluated in response to the critical need for better risk stratification. These models range from traditional scoring systems based on readily available clinical and laboratory parameters to more sophisticated approaches incorporating psychometric tests, radionics assessments, and advanced statistical techniques, including machine learning and artificial intelligence[ 9 – 12 ]. Although there are increasing risk prediction models for HE in patients with cirrhosis, the quality and applicability of these models have not been formally and systematically reviewed. Therefore, this study aims to identify, appraise, and synthesize the evidence on the performance and predictors of existing prediction models for the risk of developing HE in patients with cirrhosis. The findings will provide valuable references for clinical practice and study design for future HE prediction. Methods The protocol for this systematic review and meta-analysis was registered on PROSPERO (CRD420251040913). Search strategy To conduct a comprehensive search, both Chinese and English databases were considered. The electronic databases searched included China National Knowledge Infrastructure (CNKI), WanFang Database, China Science and Technology Journal Database (VIP), SinoMed, PubMed, Embase, Web of Science, Scopus, and the Cochrane Library, from their establishment until April 12, 2025. Databases were searched using Medical Subject Headings (MeSH) and entry terms. No further search constraints were implemented. The search terms included “Liver Cirrhosis”, “Cirrhosis”, “Liver Fibrosis”, “Hepatic Encephalopathy”, “Hepatic Coma”, “Risk prediction model”, “Risk factor”, “Predictor”, “Model”, “Risk score”, and “Prediction tool”. The specific search strategy is shown in the Supplemental Materials. Furthermore, we manually searched references in the pertinent literature for additional potentially suitable studies. We employed the PICOTS framework for the systematic review, recommended by the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist[ 13 ]. This system helps frame the review's aim, search strategy, and study inclusion and exclusion criteria. The key items of the systematic review are described below: P (Population): Patients with cirrhosis (whether or not TIPS treatment is accepted). I (Intervention): Any prognostic model developed and published to predict HE risk in patients with cirrhosis. C (Comparator): No competing model. O (Outcome): The outcome focused on the occurrence of HE rather than other liver-related events and survival rate. T (Timing): The outcome was predicted after evaluating basic information, laboratory indicators, genetic and imaging characteristics, and other factors without imposing any specific restrictions on the scope of the prediction. S (Setting): The intended role of the risk prediction model is to individualize the prediction of the probability of developing HE in patients with cirrhosis to guide risk mitigation strategies. Inclusion and exclusion criteria The inclusion criteria for studies were: (1) patients aged ≥ 18 years who met the liver cirrhosis diagnostic criteria. (2) the types of studies included case-control studies, cohort studies, and cross-sectional studies. (3) a prognostic prediction model was constructed. (4) the outcome of interest was HE, including the minimal, covert, and overt types. (5) articles written in English or Chinese. The exclusion criteria for studies were: (1) focusing only on risk factors without constructing a prediction model. (2) diagnostic model or non-original model. (3) HE data not reported separately. (4) the full text of the literature is not available or unable to extract data. (5) review, meta-analysis, case report, conference paper, dissertation, and other types of literature. Literature selection and screening Two researchers (ZX and XL) independently conducted the literature selection process. Initially, duplicate literature was identified utilizing EndNote 20 software and removed manually. Subsequently, the remaining studies were further evaluated according to their titles and abstracts to determine eligibility. Ultimately, following the application of the inclusion and exclusion criteria, the full text was reviewed to determine the studies included. Disagreements that emerged throughout the screening process were resolved through consultation and discussion with a third researcher (FZ). Data extraction Two researchers (ZX and XL) independently extracted the data, and any disagreements were resolved through consultation and discussion with a third researcher (FZ). The information extracted from all eligible articles was categorized into three groups: (1) Basic characteristics of the study: include the first author, publication year, cohort source, study design, main outcome, sample size, incidence of HE, age, gender distribution, TIPS status, research period, etiology, and patient type. (2) Basic information of the models: include the variable selection method, model development method, calibration method, missing data handling, validation method, final predictors, model presentation, and model performance. Potential indicators for assessing model calibration included the Brier score, calibration plot, calibration curve, and the Hosmer-Lemeshow test. We extracted data for each model separately if multiple models were described in the same article. (3) Distribution of predictors in the model: include the general information, laboratory indicators, comorbidities, imaging features, genetic features, Chinese medicine symptoms, and psychological tests. Risk of bias assessment To evaluate the risk of bias (ROB) and applicability of the included studies, we employed the available version of the Prediction Model Risk of Bias Assessment Tool (PROBAST)[ 14 , 15 ]. PROBAST enables a focused and transparent approach to assessing the ROB and applicability of studies that develop, validate, or update prediction models for individualized predictions[ 14 ]. PROBAST is organized into the following four domains: participants, predictors, outcome, and analysis, which contain a total of 20 signaling questions. Each signaling question can be answered as “yes/probably yes”, “no/probably no”, or “no information”. If at least one signaling question in a domain is answered with “no/probably no”, that domain should be considered at high risk of bias. Overall bias can only be deemed low risk when all domains are evaluated as low risk of bias. Two researchers (ZX and XL) independently employed the PROBAST tool for assessment, and any disagreements were resolved through consultation and discussion with a third researcher (FZ). Statistical analysis We employed Stata software (version 17.0) to conduct a meta-analysis of the area under the curve (AUC) for the externally validated prediction models. For studies reporting the c-statistic without specifying AUC, we treated the c-statistic as equivalent to AUC when the prediction model was designed for binary outcomes[ 16 ]. Meta-analysis of predictors was performed using Review Manager 5.4 software, and the combined odds ratio (OR) value and its 95% confidence intervals (CI) were calculated using the OR value as the effect indicator. Due to the large variety of predictors and the few studies reporting OR values, a meta-analysis was performed on only the predictors with the top ten ranked probabilities. In principle, the predictors with a combined number of studies of two or more were included. The I 2 statistic and the Cochrane’s Q test were used to evaluate the extent of heterogeneity. I 2 values ≤ 25% indicate low heterogeneity, 25% 50% indicate high heterogeneity. If P 50%, the random effects model was utilized; otherwise, heterogeneity was deemed acceptable, and the fixed effects model was chosen. In addition, the AUC value measures the effect size: 0.5–0.7 indicates poor predictive performance, 0.7–0.9 indicates moderate predictive performance, and 0.9-1.0 indicates excellent predictive performance. Finally, Egger’s test was used to assess publication bias. The Egger’s test with a p-value < 0.05 was considered publication bias. Results Literature selection Figure 1 shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-2020) flowchart, illustrating the exhaustive search process and results. A total of 14,728 literature records were initially retrieved. After removing 5,889 duplicate records, 8,839 studies were screened for titles and abstracts. Following the preliminary screening, 94 articles were included for full-text assessment, resulting in the exclusion of 56 studies, one of which included a non-cirrhosis population, ten focused on incorrect outcomes, seven were not available in full text or unable to extract data, 27 reported diagnostic models or non-original models, and 11 republished the same content. The Supplementary Material describes information about studies excluded after assessing the full text. Ultimately, 38 best prediction models from 38 studies were included in this review[ 8 – 12 , 17 – 49 ]. Study characteristics Table 1 summarizes the specific characteristics of the 38 included studies. These studies were published between 2015 and 2025, of which 30 were conducted in China (13 studies published in Chinese[ 37 – 49 ]), four in America, three in Europe, and one in Japan. Of the included studies, six were prospective (including four multicenter studies), and 32 were retrospective, of which 26 were conducted in a single center. In terms of study outcomes, 17 studies focused on the development of OHE, four on the development of CHE and MHE, the remaining studies were unspecified, and 17 specifically addressed HE after TIPS. The etiology of cirrhosis in the studies was broad, including alcohol, viral, non-alcoholic fatty liver disease (NAFLD), autoimmune, cholestatic, and other, with three studies all focusing on patients with hepatitis B, and five studies did not specify the etiology of cirrhosis. The sample sizes of the included studies ranged from 108 to 1,979, and the incidence of HE ranged from 7.2% to 50.4%. Models information Table 2 presents detailed information about the models in the included studies. Among the included studies, the most common modeling method used was logistic regression analysis (26 studies), followed by machine learning (8 studies) and Cox regression analysis (5 studies). Among the machine learning methods, the most common was support vector machine (SVM, seven studies), and the rest included random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), decision tree, and artificial neural network (ANN). In addition, nine studies used two or more modeling methods. Eighteen studies reported calibration methods (including calibration plots, calibration curves, Brier score, and the Hosmer-Lemeshow test), and five studies reported the handling of missing data. Among the included studies, the majority of the models were internally or externally validated. Of these, 25 studies were internally validated, 15 were externally validated, and 10 were both internally and externally validated, while eight studies did not undergo any validation after development. The primary method of model presentation involved nomograms, including 14 studies. The reported AUC or c-statistic values in the models ranged from 0.667 to 0.969. Predictors in the model Figure 2 shows the distribution of predictors in the models. Among the included models, the most commonly used predictors were age and Child-Pugh grade/score, which appeared in 16 and 12 models, respectively. Other commonly used predictors included imaging characteristics, albumin, total bilirubin, previous HE, and creatinine. Figure 3 illustrates at least 10 predictors by predictor category. Risk of bias assessment Table 3 and Fig. 4 summarize the risk of bias and applicability of the included studies. Thirty-four studies were found to have a high risk of bias, and 27 studies raised applicability concerns. In the participant domain, four studies were considered to have a high risk of bias, primarily due to the use of inappropriate data sources, with the included studies being retrospective non-cohort studies[ 37 , 38 , 43 , 45 ]. In the predictor domain, eight studies were identified as having a high risk of bias due to potential unblinding in the assessment of predictors[ 11 , 20 , 25 , 30 , 37 , 39 , 43 , 47 ]. In the outcome domain, six studies were found to be at a high risk of bias, while eleven studies were classified as having an unclear risk of bias due to the lack of blinded assessments between outcome and predictor or potential unblinding[ 11 , 12 , 18 , 20 , 24 – 27 , 29 – 31 , 34 , 36 – 38 , 41 , 43 ]. In the analysis domain, the majority of studies containing 31 items were found to be at high risk of bias, and three studies had an uncertain risk of bias. Of them, four studies had insufficient sample sizes of less than 20 events per variable (EPV)[ 10 – 12 , 46 ]; three did not include all participants[ 22 , 34 , 38 ]; twenty-two did not avoid selecting predictors based on univariate analysis[ 9 , 11 , 19 , 21 , 24 , 25 , 29 , 32 , 34 , 36 , 37 , 39 – 49 ]; eighteen did not account for model overfitting and optimism in model performance[ 8 , 10 , 19 , 24 – 27 , 29 , 30 , 35 – 37 , 42 – 45 , 48 , 49 ]. In terms of applicability assessment, 27 studies were classified as high risk, and 11 studies were classified as low risk. In the participant domain, 26 studies were deemed to have a high applicability risk, as some studies solely included participants with TIPS or non-TIPS. In the predictor domain, three studies were considered high risk for applicability because there were concerns about predictor assessment[ 8 , 22 , 47 ]. In the outcome domain, high concerns about applicability existed for three studies because one study focused only on advanced HE and two studies focused only on HE within three months[ 8 , 24 , 31 ]. Meta-analysis Among the studies included, only eight reported the model’s performance (AUC or c-statistic) along with its 95% CI after external validation. Of them, two studies reported the c-statistic based on survival analysis models[ 17 , 21 ], four studies developed models using logistic regression and reported AUC[ 8 , 31 , 41 , 47 ], one study used a machine learning model[ 38 ], and one study used a Fine-Gray competing risk model[ 27 ]. Consequently, only four studies that also employed logistic regression models were incorporated into the meta-analysis. The combined AUC value based on the fixed-effects model was 0.802 (95% CI: 0.785–0.820) (Fig. 5 ). The I 2 value was 0% (P = 0.658), indicating low heterogeneity between studies, and the P-value for Egger’s test was 0.193, suggesting no significant publication bias. Of at least 70 predictors, only 10 included ≥ 5 studies and were thus classified as major predictors for inclusion in the meta-analysis. However, a meta-analysis of the MELD score and creatinine could not be performed due to the number of studies reporting OR values being fewer than two. The meta-analysis showed that age (OR = 1.04, 95% CI: 1.03, 1.05), prior HE (OR = 4.42, 95% CI: 2.67, 7.31), low albumin (OR = 1.78, 95% CI: 1.25, 2.56), total bilirubin (OR = 2.22, 95% CI: 1.73, 2.85), Child-Pugh grade/score (OR = 2.41, 95% CI: 1.87, 3.09; OR = 1.65, 95% CI: 1.16, 2.33, respectively), ascites (OR = 1.96, 95% CI: 1.48, 2.60), and co-infection (OR = 2.57, 95% CI: 1.66, 3.98) were significant predictors of HE in cirrhosis (P < 0.01). The OR for serum sodium was 0.92 (95% CI: 0.80–1.05) (P = 0.21), indicating a not statistically significant difference. However, the analysis included only two studies, and the heterogeneity was high (I 2 = 67%, P = 0.08), which requires a cautious interpretation of the results. In addition, the heterogeneity of Child-Pugh score was also high (I 2 = 90%, P < 0.001). The results of the Egger’s test indicated the presence of publication bias in studies about Child-Pugh grade and ascites, with P-values of 0.002 and 0.026, respectively. Table 4 and Fig. 6 present the results of the meta-analysis of risk factors for HE in patients with cirrhosis. Discussion HE is a common and serious complication in patients with cirrhosis, occurring in 30% to 45% of cases, and is an independent risk factor for death in cirrhotic patients[ 50 , 51 ]. Accurately predicting HE risk is crucial for guiding clinical decisions, optimizing resource allocation, and improving patient outcomes. Recently, prediction models based on clinical, radiomic, and psychological tests have emerged as key tools for the early identification of HE. This review aimed to comprehensively assess the performance of existing models for predicting HE risk in patients with cirrhosis and to summarize their value for clinical application, although most of these are based on Chinese patient data. This study included 38 studies, each reporting a model with the best predictive performance involving AUC values ranging from 0.667 to 0.969. However, according to the PROBAST checklist, most studies had a high risk of bias and applicability concerns, which may be related to the study design, patient population characteristics, definition of HE, and insufficient model validation. The combined AUC value of the models included in the meta-analysis was 0.802 (95% CI: 0.785–0.820); however, the small number of studies involved may affect the extrapolation of the results. Furthermore, several studies failed to adequately adhere to the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) statement[ 52 ]. Specifically, the methods for variable selection, handling of missing data, calibration methods, model performance metrics, and confidence intervals were not sufficiently detailed, potentially introducing bias and uncertainty into the models. Logistic regression is a widely used modeling method, with 26 studies in this review employing logistic regression analysis, followed by 8 studies utilizing machine learning. While machine learning models have the potential to enhance predictive power and accuracy over traditional logistic regression models[ 53 , 54 ], a systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models[ 55 ]. Additionally, machine learning models often lack suitable representation tools, which may hinder their further clinical application[ 56 ]. Given the current evidence, it remains challenging to determine which model is superior for predicting the risk of HE when comparing prediction models developed using various methods. Therefore, in the future development and application of predictive models, it is crucial to thoroughly consider the model’s performance characteristics, the relevant population, and its practicality in clinical scenarios. In the prediction models reported in this review, common predictors included age, previous HE, albumin, total bilirubin (TBIL), Child-Pugh grade/score, MELD score, serum sodium, creatinine, ascites, and infection. These predictors are easily accessible and cost-effective, making them valuable for future research and clinical practice. Firstly, age serves as an independent risk factor for HE in individuals with cirrhosis. Older patients often experience sarcopenia, which can reduce the ability of muscle mitochondria to detoxify ammonia, and ageing itself can negatively affect brain function, further increasing the risk of HE[ 57 , 58 ]. Secondly, patients with a history of OHE episodes have a 42% risk of recurrence within one year[ 5 ]. The persistence of cognitive dysfunction and porto-systemic shunting diminishes the body’s capacity to compensate for these issues, highlighting the need for secondary prevention and health education for those with a previous HE history. Moreover, traditional scoring systems, such as Child-Pugh and MELD, are frequently employed to estimate the risk of HE, as they partially indicate the extent of liver failure. The Child-Pugh grading system considers the patient’s overall condition, ascites, TBIL, albumin levels, and prothrombin time. In contrast, the MELD score incorporates TBIL, creatinine levels, international normalized ratio (INR), and cirrhosis etiology to evaluate the liver functional reserve and prognosis for patients with chronic liver disease. Both scoring systems include TBIL, indicating its significance in predicting liver function. Research suggests that TBIL is an independent factor influencing mortality among HE patients[ 59 ]. Elevated levels of bilirubin, particularly unconjugated bilirubin, can contribute to or worsen the development of HE through various mechanisms, including direct neurotoxicity, exacerbation of oxidative stress and inflammation, impairment of mitochondrial function, and effects on blood-brain barrier permeability[ 60 ]. Albumin is often used to assess nutritional status, and patients with decompensated cirrhosis are usually associated with varying degrees of malnutrition and sarcopenia, which can contribute to the conversion of gluconeogenesis and increase ammonia production when the body is in a state of negative nitrogen balance[ 61 ]. Studies have indicated that albumin is an independent risk factor for the development of HE and that albumin infusion reduces the incidence and severity of HE in patients with cirrhosis[ 62 , 63 ]. In addition, a prospective study showed that the presence of hyponatremia is a major risk factor for the development of OHE in cirrhotic patients[ 64 ]. The study suggests that developing hyponatremia in the context of disturbed osmotic homeostasis in the cirrhotic brain may represent a second osmotic blow to astrocytes[ 64 ]. The OR value for serum sodium combined in this meta-analysis was 0.92 (95% CI: 0.80–1.05), a difference that was not statistically significant, which is not in line with previous studies, mainly due to biased results from the small number of included studies. Elevated serum creatinine levels usually reflect impaired renal function, which exacerbates the accumulation of toxins in the body and systemic inflammation, thereby contributes to the development of HE. Including creatinine in the MELD score also signals the important role of renal factors in the prognosis of liver cirrhosis. Similarly, the development of ascites in cirrhotic patients is closely associated with hepatic decompensation and exacerbation of portal hypertension, which is one of the key drivers of hepatic encephalopathy[ 65 ]. Finally, infection is a recognized HE trigger in cirrhotic patients, most commonly seen in spontaneous bacterial peritonitis, associated with hyperammonemia due to intestinal bacterial overgrowth[ 66 ]. At the same time, infections can further worsen astrocyte dysfunction already caused by hyperammonemia through the direct action of inflammatory mediators, cerebral microvascularisation and endothelial metabolism[ 67 , 68 ]. In addition to traditional clinical and laboratory indicators, emerging models have started to incorporate radiomic features, genetic factors, and psychological assessments into their prediction systems, enhancing their ability to predict the occurrence of HE. Among the studies examined, the radiomic factors included computed tomography (CT) features related to visceral adipose tissue, the liver, the spleen, and the liver-related vascular morphology[ 11 , 18 – 20 , 25 , 30 ]. However, it is important to note that obtaining these factors through imaging techniques can be cumbersome and costly, which means their use in clinical practice requires further evaluation. Genetic risk factors for HE have been previously identified, and cirrhotic patients carrying GLS (encoding glutaminase) variants will have a significantly increased incidence of HE[ 7 ]. Integrating genetic background information into prediction models may prove beneficial for forecasting HE in cirrhotic patients, but further validation in diverse populations with varying genetic backgrounds is essential[ 21 ]. It is well established that the psychometric hepatic encephalopathy score (PHES) is the gold standard for diagnosing MHE[ 69 ]. Recent studies have confirmed that the PHES can predict the development of OHE and mortality in cirrhotic patients[ 70 ]. However, psychometric tests can be highly subjective, reliant on the patient, and time-consuming. Future research should explore how these tests can be simplified and integrated into predictive models. Our study has several potential limitations. First, most of the included studies were conducted in China, which may limit the generalizability of the findings to other populations. This suggests that adjustments may be needed when applying these models in different regions, highlighting the need to develop risk prediction models in more diverse populations in the future. Second, only four studies that were externally validated were ultimately included in the meta-analysis due to insufficient information reported by some of the studies. This could lead to bias in assessing the heterogeneity between studies and may reduce the statistical efficacy related to publication bias. Additionally, there may be differences in the objectivity and consistency of diagnosing or staging HE across studies, which can impact the pooled results. However, these issues do not affect the systematic evaluation of prediction models. Finally, we were unable to conduct subgroup analyses for all potential confounders due to limitations in data availability. Conclusions This systematic review includes thirty-eight studies and models. The four externally validated models achieved a combined AUC of 0.802 (95% CI: 0.785–0.820), indicating moderate predictive performance. This study showed that age, prior HE, low albumin, total bilirubin, Child-Pugh grade/score, ascites, and co-infection were significant predictors of HE in cirrhosis. However, according to the PROBAST checklist, thirty-four studies were found to have a high risk of bias, and twenty-seven studies raised concerns regarding their applicability. Therefore, the clinical utility of the model requires further validation. Future research should focus on developing new models through optimised study design and analysis, increased sample sizes and external validation, and applying them to clinical practice. Abbreviations HE Hepatic encephalopathy CHE Covert hepatic encephalopathy MHE Minimal hepatic encephalopathy OHE Overt hepatic encephalopathy MELD Model for end-stage liver disease TIPS Transjugular intrahepatic portosystemic shunt CNKI China National Knowledge Infrastructure VIP China Science and Technology Journal Database MeSH Medical Subject Headings CHARMS Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies ROB Risk of bias PROBAST Prediction Model Risk of Bias Assessment Tool AUC Area under the curve OR Odds ratio CI Confidence intervals PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses NAFLD Non-alcoholic fatty liver disease SVM Support vector machine RF Random forest XGBoost Extreme gradient boosting LightGBM Light gradient boosting machine ANN Artificial neural network TRIPOD Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis TBIL Total bilirubin INR International normalized ratio CT Computed tomography PHES Psychometric hepatic encephalopathy score Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Availability of data and materials All data generated or analysed during this study are included in this published article and its supplementary information files. Competing interests The authors declare that they have no competing interests. Funding This work was supported by National Natural Science Foundation of China (grant number 82070569). Authors’ contributions Zhi-Le Xiao: Conceptualization and methodology, Data curation, Formal analysis, Visualization and validation, Original draft writing. Xiao-Nan Li: Data curation, Formal analysis, Visualization and validation. Feng Zhou: Conceptualization and methodology, Data curation, Formal analysis, Visualization and validation, Project administration. All authors read and approved the final manuscript. Acknowledgments Not applicable. References Häussinger D, Dhiman RK, Felipo V, et al. Hepatic encephalopathy. Nat Rev Dis Primer. 2022;8:43. Montagnese S, Rautou P-E, Romero-Gómez M, Larsen FS, Shawcross DL, Thabut D, Vilstrup H, Weissenborn K. EASL Clinical Practice Guidelines on the management of hepatic encephalopathy. J Hepatol. 2022;77:807–24. Louissaint J, Deutsch-Link S, Tapper EB. Changing Epidemiology of Cirrhosis and Hepatic Encephalopathy. Clin Gastroenterol Hepatol. 2022;20:S1–8. Vidal-Cevallos P, Chávez-Tapia NC, Uribe M. Current approaches to hepatic encephalopathy. Ann Hepatol. 2022;27:100757. Rose CF, Amodio P, Bajaj JS, Dhiman RK, Montagnese S, Taylor-Robinson SD, Vilstrup H, Jalan R. Hepatic encephalopathy: Novel insights into classification, pathophysiology and therapy. J Hepatol. 2020;73:1526–47. Nardelli S, Lattanzi B, Merli M, Farcomeni A, Gioia S, Ridola L, Riggio O. Muscle Alterations Are Associated With Minimal and Overt Hepatic Encephalopathy in Patients With Liver Cirrhosis. Hepatology. 2019;70:1704–13. Mayer LB, Krawczyk M, Grünhage F, Lammert F, Stokes CS. A genetic variant in the promoter of phosphate-activated glutaminase is associated with hepatic encephalopathy. J Intern Med. 2015;278:313–22. Bajaj JS, Tandon P, O’Leary JG, et al. Admission Serum Metabolites and Thyroxine Predict Advanced Hepatic Encephalopathy in a Multicenter Inpatient Cirrhosis Cohort. Clin Gastroenterol Hepatol. 2023;21:1031–e10403. Tapper EB, Parikh ND, Sengupta N, Mellinger J, Ratz D, Lok AS-F, Su GL. A risk score to predict the development of hepatic encephalopathy in a population‐based cohort of patients with cirrhosis. Hepatology. 2018;68:1498–507. Yi C, Wei Y, Zhengcui H, Ying X, Xiaoying L. Construction of an early warning model of hepatic encephalopathy in patients with liver cirrhosis based on psychological testing. J Clin Nurs. 2023;32:191–8. Cheng S, Hu G, Jin Z, Wang Z, Xue H. Prediction of Hepatic Encephalopathy After Transjugular Intrahepatic Portosystemic Shunt Based on CT Radiomic Features of Visceral Adipose Tissue. Acad Radiol. 2024;31:1849–61. Yang H, Li X, Cao H, Cui Y, Luo Y, Liu J, Zhang Y. Using machine learning methods to predict hepatic encephalopathy in cirrhotic patients with unbalanced data. Comput Methods Programs Biomed. 2021;211:106420. Moons KGM, de Groot JAH, Bouwmeester W, Vergouwe Y, Mallett S, Altman DG, Reitsma JB, Collins GS. Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS Checklist. PLoS Med. 2014;11:e1001744. Wolff RF, Moons KGM, Riley RD, Whiting PF, Westwood M, Collins GS, Reitsma JB, Kleijnen J, Mallett S. PROBAST: A Tool to Assess the Risk of Bias and Applicability of Prediction Model Studies. Ann Intern Med. 2019;170:51. Moons KGM, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, Reitsma JB, Kleijnen J, Mallett S. PROBAST: A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies: Explanation and Elaboration. Ann Intern Med. 2019;170:W1. Verbakel JY, Steyerberg EW, Uno H, De Cock B, Wynants L, Collins GS, Van Calster B. ROC curves for clinical prediction models part 1. ROC plots showed no added value above the AUC when evaluating the performance of clinical prediction models. J Clin Epidemiol. 2020;126:207–16. Ballester MP, Tranah TH, Balcar L, et al. Development and validation of the AMMON-OHE model to predict risk of overt hepatic encephalopathy occurrence in outpatients with cirrhosis. J Hepatol. 2023;79:967–76. Cao J, Yang J, Ming Z, et al. A radiomics model of liver CT to predict risk of hepatic encephalopathy secondary to hepatitis B related cirrhosis. Eur J Radiol. 2020;130:109201. Chen X, Huang M, Yu X, et al. Hepatic-associated vascular morphological assessment to predict overt hepatic encephalopathy before TIPS: a multicenter study. Hepatol Int. 2024;18:1238–48. Chen X, Wang T, Ji Z, et al. 3D automatic liver and spleen assessment in predicting overt hepatic encephalopathy before TIPS: a multi-center study. Hepatol Int. 2023;17:1545–56. Gil-Gómez A, Ampuero J, Rojas Á, et al. Development and Validation of a Clinical-Genetic Risk Score to Predict Hepatic Encephalopathy in Patients With Liver Cirrhosis. Am J Gastroenterol. 2021;116:1238–47. İnce O, Önder H, Gençtürk M, Golzarian J, Young S. Machine Learning Insights: Predicting Hepatic Encephalopathy After TIPS Placement. Cardiovasc Intervent Radiol. 2023;46:1715–25. Li K, Cheng Y, Zhao R, Jiang H, Zhang L, Tong Y, Li S. Prediction of mortality and overt hepatic encephalopathy undergoing transjugular intrahepatic portosystemic shunt: a retrospective cohort study. Abdom Radiol. 2024;49:908–18. Li Y, He X, Pang H. A model to predict early hepatic encephalopathy in patients undergoing transjugular intrahepatic portosystemic shunt. Turk J Gastroenterol. 2019;30:702–7. Liu J, Zhou C, Wang Y, Yang C, Shi Q, Huang S, Chen Y, Li T, Xiong B. The combination of Child–Pugh score and quantitative CT-based spleen volume could predict the risk of hepatic encephalopathy after transjugular intrahepatic portosystemic shunt creation. Abdom Radiol. 2021;46:3464–70. Miwa T, Hanai T, Nishimura K, Maeda T, Tajirika S, Imai K, Suetsugu A, Takai K, Yamamoto M, Shimizu M. A simple covert hepatic encephalopathy screening model based on blood biochemical parameters in patients with cirrhosis. PLoS ONE. 2022;17:e0277829. Riggio O, Amodio P, Farcomeni A, et al. A Model for Predicting Development of Overt Hepatic Encephalopathy in Patients With Cirrhosis. Clin Gastroenterol Hepatol. 2015;13:1346–52. Tapper EB, Zhao L, Nikirk S, Baki J, Parikh ND, Lok AS, Waljee AK. Incidence and Bedside Predictors of the First Episode of Overt Hepatic Encephalopathy in Patients With Cirrhosis. Am J Gastroenterol. 2020;115:2017–25. Tong H, Gan C, Wei B, et al. Risk factors for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt creation in patients with liver cirrhosis. J Dig Dis. 2021;22:31–40. Wang C, Teng Y, Gao J, Zhang Z, Li Y. Low adipose tissue index as an indicator of hepatic encephalopathy in cirrhotic patients following transjugular intrahepatic portosystemic shunt. Abdom Radiol. 2023;48:1454–67. Yang C, Zhu X, Liu J, et al. Development and Validation of Prognostic Model to Estimate the Risk of Overt Hepatic Encephalopathy After TIPS Creation: A Multicenter Study. Clin Transl Gastroenterol. 2022;13:e00461. Yang Y, Fu S, Cao B, et al. Prediction of overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt treatment: a cohort study. Hepatol Int. 2021;15:730–40. Yang Y, Liang X, Yang S, et al. Preoperative prediction of overt hepatic encephalopathy caused by transjugular intrahepatic portosystemic shunt. Eur J Radiol. 2022;154:110384. Yao C, Huang L, Wang M, et al. Establishment and validation of a nomogram model for riskprediction of hepatic encephalopathy: a retrospective analysis. Sci Rep. 2023;13:19544. Zeng X, Yin C, Sun CY, et al. Prevalence and risk factors of covert hepatic encephalopathy in cirrhotic patients: A multicenter study in China. J Dig Dis. 2023;24:122–32. Zhong B-Y, Wang W-S, Shen J, Du H, Zhang S, Li W-C, Yin Y, Yang J, Ni C-F, Zhu X-L. Single-Centre Retrospective Training Cohort Using Artificial Intelligence for Prognostic Prediction of Encephalopathy, Mortality, and Liver Dysfunction after Early TIPS Creation. Cardiovasc Intervent Radiol. 2021;44:1597–608. Chen L, Pan C, Liu M, Yi L. Construction of risk assessment model for hepatic encephalopathy in patients with liver cirrhosis and its predictive value. China Mod Dr. 2023;61:29–3273. Fu S, He B, Xu J. Establishment and validation of prediction model for cirrhosis-related hepatic encephalopathy by machine learning algorithm. Chin J Lab Med. 2025;48:93–102. Gao X, Peng F, Li P. Analysis of risk factors of liver cirrhosis complicated with mild hepatic encephalopathy and construction of nomogram prediction model. Chin J Integr Tradit West Med Liver Dis. 2024;34:9–15. Huang T, Wang G, Huang G, Zhang C. Prediction of different clinical outcomes in patients with cirrhosis after TIPS based on machine learning algorithm. Chin J Dig Med Imageology Electron Ed. 2022;12:4–10. Li XQ, Li Y, Ni YQ, Cao W, Yin TT, Lu R. A nomogram prediction model for individualized prediction of the risk of covert (minimal) hepatic encephalopathy occurrence in patients with liver cirrhosis. Zhonghua Gan Zang Bing Za Zhi. 2024;32:828–34. Li X, Li J, Sun W. Establishment and evaluation of a nomogram prediction model for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt in patients with liver cirrhosis and esophagogastric variceal bleeding. J Clin Hepatol. 2024;40:1605–11. Tan JT, Xu XM, He YX, Chao T, Jun G, Liu YY, Xiang SS, Zhao WL. Construction of prediction model of cirrhosis-related hepatic encephalopathy based on machine learning algorithm. Med J Chin Peoples Lib Army. 2021;46:354–60. Wang L, Qin J, Yao X, Qi Q, Liu L, Tang S. Construction of a risk prediction model for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt in patients with hepatitis B cirrhosis and portal hypertension. J Clin Hepatol. 2024;40:1149–55. Wang N, Li J, Li X, Liang L, Gan X, Wang X. The development and application of a risk prediction model for hepatic encephalopathy in patients with liver cirrhosis. Chin J Nurs. 2019;54:805. Wang X, Zhai M, Ren H, Li M, Quan D, Zhang Y, Liu J, Qiu L. An exploratory study of a classification model based on resampling and Voting heterogeneous integration in the prediction of the risk of liver cirrhosis complicated by hepatic encephalopathy. Chin J Health Stat. 2022;39:545–9. Wen J, Zhao Y, Wang S, et al. Construction and validation of a risk prediction model for liver cirrhosis complicated with minimal hepatic encephalopathy based on the involvement of traditional Chinese medicine factors. Chin J Liver Dis Electron Version. 2024;16:48–56. Yuan W, Li L, Han X, Liu J, Li B, Wu J. Clinical prediction modeling and efficacy analysis of the occurrence of hepatic encephalopathy after TIPS. Mod Dig Interv. 2021;26:874–7. Zhang J, Wang Z, Yang J, Wang X, Zhao Y. Construction and validation of the risk prediction model for secondary hepatic encephalopathy in patients with chronic hepatitis B and cirrhosis. Chin Hepatol. 2023;28:916–20. Elwir S, Rahimi RS. Hepatic Encephalopathy: An Update on the Pathophysiology and Therapeutic Options. J Clin Transl Hepatol. 2017;5:142–51. Bajaj JS, O’Leary JG, Tandon P, et al. Hepatic Encephalopathy Is Associated With Mortality in Patients With Cirrhosis Independent of Other Extrahepatic Organ Failures. Clin Gastroenterol Hepatol. 2017;15:565–e5744. Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594–7594. Rajkomar A, Dean J, Kohane I. Machine Learning in Medicine. N Engl J Med. 2019;380:1347–58. Churpek MM, Yuen TC, Winslow C, Meltzer DO, Kattan MW, Edelson DP. Multicenter Comparison of Machine Learning Methods and Conventional Regression for Predicting Clinical Deterioration on the Wards. Crit Care Med. 2016;44:368–74. Christodoulou E, Ma J, Collins GS, Steyerberg EW, Verbakel JY, Van Calster B. A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models. J Clin Epidemiol. 2019;110:12–22. Fu H, Hou D, Xu R, You Q, Li H, Yang Q, Wang H, Gao J, Bai D. Risk prediction models for deep venous thrombosis in patients with acute stroke: A systematic review and meta-analysis. Int J Nurs Stud. 2024;149:104623. Welch N, Attaway A, Bellar A, Alkhafaji H, Vural A, Dasarathy S. Compound Sarcopenia in Hospitalized Patients with Cirrhosis Worsens Outcomes with Increasing Age. Nutrients. 2021;13:659. Xu X, Ding H, Li W, et al. Chinese Guidelines on the Management of Hepatic Encephalopathy in Cirrhosis (2024). J Clin Transl Hepatol. 2025;000:000–000. Cui Y, Guan S, Ding J, He Y, Li Q, Wang S, Sun H. Establishment and evaluation of a model for predicting 3-month mortality in Chinese patients with hepatic encephalopathy. Metab Brain Dis. 2019;34:213–21. Pranty AI, Shumka S, Adjaye J. Bilirubin-Induced Neurological Damage: Current and Emerging iPSC-Derived Brain Organoid Models. Cells. 2022;11:2647. Amodio P, Bemeur C, Butterworth R, Cordoba J, Kato A, Montagnese S, Uribe M, Vilstrup H, Morgan MY. The nutritional management of hepatic encephalopathy in patients with cirrhosis: International society for hepatic encephalopathy and nitrogen metabolism consensus. Hepatology. 2013;58:325–36. Bai Z, Guo X, Tacke F, Li Y, Li H, Qi X. Association of serum albumin level with incidence and mortality of overt hepatic encephalopathy in cirrhosis during hospitalization. Ther Adv Gastroenterol. 2019;12:1756284819881302. Bai Z, Bernardi M, Yoshida EM, Li H, Guo X, Méndez-Sánchez N, Li Y, Wang R, Deng J, Qi X. Albumin infusion may decrease the incidence and severity of overt hepatic encephalopathy in liver cirrhosis. Aging. 2019;11:8502–25. Guevara M, Baccaro ME, Torre A, et al. Hyponatremia Is a Risk Factor of Hepatic Encephalopathy in Patients With Cirrhosis: A Prospective Study With Time-Dependent Analysis. Am J Gastroenterol. 2009;104:1382–9. Thomsen KL, Sørensen M, Kjærgaard K, Eriksen PL, Lauridsen MM, Vilstrup H. Cerebral Aspects of Portal Hypertension. Clin Liver Dis. 2024;28:541–54. Piotrowski D, Boroń-Kaczmarska A. Bacterial infections and hepatic encephalopathy in liver cirrhosis–prophylaxis and treatment. Adv Med Sci. 2017;62:345–56. Guimarães L, Piedade J, Duarte J, Baldin C, Victor L, Costa B, Veiga Z, Alcântara C, Fernandes F, Pereira G. Hepatic Encephalopathy in Cirrhotic Patients With Bacterial Infections: Frequency, Clinical Characteristics, and Prognostic Relevance. J Clin Exp Hepatol. 2023;13:559–67. Jaffe A, Lim JK, Jakab SS. Pathophysiology of Hepatic Encephalopathy. Clin Liver Dis. 2020;24:175–88. American Association for the Study of Liver Diseases, European Association for the Study of the Liver. Hepatic Encephalopathy in Chronic Liver Disease: 2014 Practice Guideline by the European Association for the Study of the Liver and the American Association for the Study of Liver Diseases. J Hepatol. 2014;61:642–59. Labenz C, Toenges G, Schattenberg JM, Nagel M, Huber Y, Marquardt JU, Labenz J, Galle PR, Wörns M-A. Outcome Prediction of Covert Hepatic Encephalopathy in Liver Cirrhosis: Comparison of Four Testing Strategies. Clin Transl Gastroenterol. 2020;11:e00172. Tables Tables 1 to 4 are available in the Supplementary Files section. Supplementary Files Tables.docx GraphicalAbstract.docx SupplementaryMaterial.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7875969","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":555520911,"identity":"690d912d-82b8-4a09-b10d-ce540ccd0317","order_by":0,"name":"Zhile Xiao","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Zhile","middleName":"","lastName":"Xiao","suffix":""},{"id":555520912,"identity":"7c8e8d27-3836-48a6-9f2c-c84df4a322b7","order_by":1,"name":"Xiaonan Li","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Xiaonan","middleName":"","lastName":"Li","suffix":""},{"id":555520913,"identity":"8e449dd9-97c0-45ac-906c-fe4b946f1657","order_by":2,"name":"Feng Zhou","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA10lEQVRIiWNgGAWjYBAC9gYgwQgk+Bl4QHxmwlp4DkC1SDaQrMXgANFa2HsPv/y5wyZx843cYxIMFdaJDexnD+DXwnMuzULyTFritht5aRIMZ9ITG3jyEvBqsZfIMTMwbDucuO12jpkEI5DRIMFjgN8W+TdmBolAlZtng7T8I0aLBI/xg4NALRukQVoaiNHCk2PG2NiWZjzj/htji4Rj6cZtPDkEtLCfMf74s81Gtr/njOGNDzXWsv3sZ/BrAQI2CTgzAcQlpB4ImD8QoWgUjIJRMApGMgAA6kFESCcnIvsAAAAASUVORK5CYII=","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":true,"prefix":"","firstName":"Feng","middleName":"","lastName":"Zhou","suffix":""}],"badges":[],"createdAt":"2025-10-16 09:55:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7875969/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7875969/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97896364,"identity":"caf2df29-6f3d-4835-884b-d490c9929ce3","added_by":"auto","created_at":"2025-12-10 15:36:25","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17902,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/767768e3c5fcafd8f7e83892.xlsx"},{"id":97745420,"identity":"66512b8d-94b8-46c9-839c-77e1fbf00d9f","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20132,"visible":true,"origin":"","legend":"","description":"","filename":"Table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/89e205e14c85098ff4a6612c.xlsx"},{"id":97745421,"identity":"6bbf265e-54d7-4443-8da8-be8279581dd3","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12621,"visible":true,"origin":"","legend":"","description":"","filename":"Table3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/f7e3bc5946c1a397be6940cb.xlsx"},{"id":97894965,"identity":"96d3cae0-78ab-41bd-aa6d-e3d2f1c096fb","added_by":"auto","created_at":"2025-12-10 15:33:17","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11335,"visible":true,"origin":"","legend":"","description":"","filename":"Table4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/83486a87de44cdfa618e89ef.xlsx"},{"id":97894962,"identity":"625ab87a-e022-4ee6-b851-9ec7b745178a","added_by":"auto","created_at":"2025-12-10 15:33:17","extension":"xml","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9374,"visible":true,"origin":"","legend":"","description":"","filename":"sysrSYSRD2501413.xml","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/d3138c44d2ba0ddb436797c8.xml"},{"id":97894950,"identity":"dbfba1cf-0e4a-4376-a62b-44cef3a5fc0f","added_by":"auto","created_at":"2025-12-10 15:33:15","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1195,"visible":true,"origin":"","legend":"","description":"","filename":"SYSRD250141323684.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/a97c1dcc19860f9228b6a92a.xml"},{"id":97896417,"identity":"973a2db5-23df-4ef5-b758-afb86b76b623","added_by":"auto","created_at":"2025-12-10 15:36:31","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":830,"visible":true,"origin":"","legend":"","description":"","filename":"SYSRD2501413Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/72ae617bc5d04e8652092219.xml"},{"id":97897222,"identity":"47fa0f3f-f4b0-4543-9c0c-a506ea41d515","added_by":"auto","created_at":"2025-12-10 15:37:35","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":155806,"visible":true,"origin":"","legend":"","description":"","filename":"SYSRD25014130enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/713f0404df65b47550dcb987.xml"},{"id":97745437,"identity":"d358724a-0440-414d-8ba7-171cc6bd51a5","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"eps","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":209,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/24e97f66b1d066b3da68ffe8.eps"},{"id":97895331,"identity":"dad00ece-6655-4937-82f2-6557f1966b2a","added_by":"auto","created_at":"2025-12-10 15:34:00","extension":"eps","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":209,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/c71951f83e3cf90381345170.eps"},{"id":97745433,"identity":"e0894004-70e7-4779-8568-299d336996a4","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"eps","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24287,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage4.eps","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/2122e199f9b1aca649a46e21.eps"},{"id":97745449,"identity":"736c8468-1dd4-48be-b1e3-ed7902cd2523","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"eps","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":209,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/88b6494b62a639510da9bed9.eps"},{"id":97745425,"identity":"e0dd5ed9-669c-48f3-97d3-f93a915a7921","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"eps","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":220,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage6.eps","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/242320e89ab840b23e6a14c8.eps"},{"id":97894656,"identity":"4fb6a24f-f4bb-4069-8c0a-75f4e8c847a1","added_by":"auto","created_at":"2025-12-10 15:32:52","extension":"jpeg","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":783968,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/c4f69a69c9d0e5c5e9a7d128.jpeg"},{"id":97745438,"identity":"de346722-76fd-435d-9a1c-0848def4d9a7","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1456193,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/6cf0a4bae098c52e6ad18ef9.jpeg"},{"id":97897019,"identity":"dd76db64-04d3-4ce9-9c14-8e1c52bbe719","added_by":"auto","created_at":"2025-12-10 15:37:21","extension":"jpeg","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122416,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/2604f75679e0e107cc23e3eb.jpeg"},{"id":97745428,"identity":"c5e7c781-83a4-467b-ae7d-4aab91926483","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"emf","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":161500,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.emf","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/e7f4f1821a7bb357161be703.emf"},{"id":97745460,"identity":"9203c125-df19-4a3e-9774-5db178157088","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"emf","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102972,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.emf","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/b7869ab54b01bf1eecc773c5.emf"},{"id":97745430,"identity":"c2052273-8df0-4377-8d2c-0a74f013ad74","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84710,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage14.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/727dac06e0cea92368884a56.jpeg"},{"id":97896039,"identity":"290ff86d-fafe-49c4-bf60-62e8b67fdb60","added_by":"auto","created_at":"2025-12-10 15:35:44","extension":"jpeg","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":689718,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage15.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/4b28083ef15609efb505ea02.jpeg"},{"id":97745441,"identity":"53281ad9-2f5a-4f51-bf0d-2020126742da","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":361651,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage16.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/6138e8e4cc8f1cc615258a82.jpeg"},{"id":97745426,"identity":"85b8e74e-841e-440c-977c-7b510e31b83c","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15849,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage17.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/b6b26065f61ad35cb63ed9fb.jpeg"},{"id":97745445,"identity":"0e281783-67ee-40b1-8c8b-fa8ed573c596","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"jpeg","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":104052,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage18.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/bdc15d4c86cb3200d7d1b45a.jpeg"},{"id":97896376,"identity":"c6235165-f204-487c-8b35-442394fe4c71","added_by":"auto","created_at":"2025-12-10 15:36:28","extension":"jpeg","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1327839,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage19.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/ccffad347f2123324fcb4f7d.jpeg"},{"id":97745436,"identity":"507ae19f-5512-40d9-91ba-6556cd91f4f3","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22946,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/a6e7f4a6187378b3e869ef36.jpeg"},{"id":97896957,"identity":"eac58d46-5bac-4005-b22f-a89210d85063","added_by":"auto","created_at":"2025-12-10 15:37:17","extension":"jpeg","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage20.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/2d36398f24de0819f782cba1.jpeg"},{"id":97745431,"identity":"f2d24014-a479-4c4b-9894-2531ce7701c8","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage20.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/6ccab57c335c4921649e56ff.jpeg"},{"id":97745470,"identity":"0ce18593-3435-4c5c-b13d-7eca168977a9","added_by":"auto","created_at":"2025-12-09 00:24:39","extension":"jpeg","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":282652,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/70fbd92cbf7af667babaabae.jpeg"},{"id":97896041,"identity":"ca0f00e2-c90f-411e-8953-ac0f67675d50","added_by":"auto","created_at":"2025-12-10 15:35:44","extension":"jpeg","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1290230,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/7c826fdf920ecb2cb56a07c7.jpeg"},{"id":97745434,"identity":"9a2a5e1b-2ddb-4d2c-a618-6baa74202060","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42988,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/ef767eb8ef5d61879c8b564d.jpeg"},{"id":97745439,"identity":"fbf6a16c-d7b8-4ad4-a1cb-0da74e9da751","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1280376,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/69066a12edfa4269f2717b0a.jpeg"},{"id":97745444,"identity":"c5f02b8f-9742-4692-b769-fb1888d2b37d","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"jpeg","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":53065,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/5aa5a443d05b640d743016dd.jpeg"},{"id":97745435,"identity":"0c5267a6-f98d-46f7-bacb-78368cee278a","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"jpeg","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102481,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/02b7527040556cd5b1c70140.jpeg"},{"id":97896280,"identity":"fbc3b580-45f6-4950-90b1-8da9a86b487d","added_by":"auto","created_at":"2025-12-10 15:36:18","extension":"png","order_by":38,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":170675,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/877683a69b97caa1554f5388.png"},{"id":97745440,"identity":"cbd14714-a7fd-4450-8b23-daa2dc5d41ec","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"png","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":240504,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/0f272ec2d59de20cd8f8c46b.png"},{"id":97895924,"identity":"2f1b84f3-bc32-4823-834d-d552137f1c7e","added_by":"auto","created_at":"2025-12-10 15:35:23","extension":"png","order_by":40,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":25010,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/f92290863d8ba92d6ac8e025.png"},{"id":97745466,"identity":"411ad8b4-6225-4f98-b74f-e7c6a671890b","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":41,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1301289,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/939ad0f2dca752f3237b348f.png"},{"id":97896650,"identity":"e07dc4fd-0263-40a1-b749-cd058d10a736","added_by":"auto","created_at":"2025-12-10 15:36:52","extension":"png","order_by":42,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1591463,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/ec92db99b6e0c073b0e10bfb.png"},{"id":97895944,"identity":"7ea347e8-0aaf-4dad-a64c-47c0bbeec6cb","added_by":"auto","created_at":"2025-12-10 15:35:26","extension":"png","order_by":43,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17593,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/842469423edc831837490039.png"},{"id":97896948,"identity":"b2020ef0-b976-464e-a9f9-e1510db49de0","added_by":"auto","created_at":"2025-12-10 15:37:16","extension":"png","order_by":44,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":110019,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/9ad19b0e0945c718643267df.png"},{"id":97745461,"identity":"a52de1d2-5b9c-4280-aadd-1105e5a97968","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":45,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":51479,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage16.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/609113facf5a878730b7f6c1.png"},{"id":97895405,"identity":"941bc0cb-d7dc-41b0-9c00-659a6ad0b48f","added_by":"auto","created_at":"2025-12-10 15:34:08","extension":"png","order_by":46,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3203,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage17.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/a467a6ce09975f7fc2d63b6a.png"},{"id":97897140,"identity":"0e677c4c-c82b-498c-96fe-a221d9eb1f62","added_by":"auto","created_at":"2025-12-10 15:37:29","extension":"png","order_by":47,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23229,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage18.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/975dc49ba071f557b01b3479.png"},{"id":97745443,"identity":"73be26eb-8744-4104-ad7a-13d5a8057b6f","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":48,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":211361,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage19.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/09066c0bf1d8d2704cb0444e.png"},{"id":97896866,"identity":"ae87099e-0014-4c01-a34f-fabc750cd57d","added_by":"auto","created_at":"2025-12-10 15:37:09","extension":"png","order_by":49,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5260,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/3c71684d422844fe27a9ee1b.png"},{"id":97745447,"identity":"86387953-4013-44cf-83d0-21204e211224","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":50,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage20.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/61666d117592d6cacafa9577.png"},{"id":97745442,"identity":"c6dc553f-6f61-43e7-99fc-0fb6d89e8b8e","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"png","order_by":51,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage20.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/e8390ad56891ae18db20aaaa.png"},{"id":97895833,"identity":"86da6fd1-ea25-4d0f-a1bb-7e108aa68b14","added_by":"auto","created_at":"2025-12-10 15:35:10","extension":"png","order_by":52,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59821,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/84b22fe60e6fa2ebacea7869.png"},{"id":97745451,"identity":"91874826-2eb7-47b5-a0cf-bbdae477e550","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":53,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":280508,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/d730baa3db4958cce2d1edc3.png"},{"id":97745458,"identity":"c707a53d-c329-487e-8a31-ac4cbfbc561b","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":54,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8829,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/0b71b5dc1137371ff9a11908.png"},{"id":97745456,"identity":"6b2a1cab-5c3b-4b7a-b59e-1cc29c81b499","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":55,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":282814,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/db8232129a994833b8a2b92f.png"},{"id":97894985,"identity":"b4331c0d-95ce-48b6-8c2a-3c23eb62d94f","added_by":"auto","created_at":"2025-12-10 15:33:18","extension":"png","order_by":56,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10513,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/848e0710bf1ff67fb51155a0.png"},{"id":97745464,"identity":"69ff401f-5637-4ba5-b985-fe10727c3d71","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"png","order_by":57,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20627,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/27a21cc211da9c9908545a70.png"},{"id":97897043,"identity":"56cdbd55-8ec9-4727-87f6-2d59e4357b75","added_by":"auto","created_at":"2025-12-10 15:37:22","extension":"xml","order_by":58,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":154154,"visible":true,"origin":"","legend":"","description":"","filename":"SYSRD25014130structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/71751239adf2142230cb0419.xml"},{"id":97745465,"identity":"66e1efd4-bc7b-4606-b8ce-88154b121ca5","added_by":"auto","created_at":"2025-12-09 00:24:38","extension":"html","order_by":59,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":167138,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/0e7da83e4c5fe1204d2120f3.html"},{"id":97896053,"identity":"a3934736-7f04-4033-8ede-b5d4953b1e39","added_by":"auto","created_at":"2025-12-10 15:35:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":235777,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePRISMA flow diagram of literature search and selection.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/1edf5beaecda69bf15876ef6.png"},{"id":97745410,"identity":"bacd91c7-d5db-4180-919f-12d8feebf92c","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":53376,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistribution of predictor variables in the model.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/f8344d5892b66c01a4c579fb.png"},{"id":97896232,"identity":"8b4c2b28-9eb6-45ac-8097-5fe92be29dce","added_by":"auto","created_at":"2025-12-10 15:36:12","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":44175,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePredictors included in at least 10 by category of predictor.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/467a61a2c3fca235a222c537.png"},{"id":97745415,"identity":"91054431-bc1e-4d4c-bcee-1c999d6419cf","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":21059,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePROBAST results of the included studies.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A). Summary of Risk of Bias assessment. \u0026nbsp;(B). Summary of Applicability assessment.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/b66d25ee40ede755bcbce1bd.png"},{"id":97895861,"identity":"c79ef713-05f4-43bd-b89b-40b02efa01b7","added_by":"auto","created_at":"2025-12-10 15:35:12","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":71675,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eForest plot of the meta-analysis of pooled AUC.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/2b27072ae9339b7d9e31b873.png"},{"id":97745423,"identity":"f382c351-5f1f-424d-9517-4b774e89180d","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":550895,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eForest plots of the meta-analysis of predictors.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/99cad4f6a9f5d00e6ef4c652.png"},{"id":101754177,"identity":"ffa22e1f-d04a-40f7-b836-80036c9e8799","added_by":"auto","created_at":"2026-02-03 10:41:50","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1537551,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/2155c470-f773-4a68-91df-be0be8b0aa46.pdf"},{"id":97895753,"identity":"dd329a61-df09-429b-a6ff-a485f71dd90f","added_by":"auto","created_at":"2025-12-10 15:34:52","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":112671,"visible":true,"origin":"","legend":"","description":"","filename":"Tables.docx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/6c9bc1d433d35171b7db52c1.docx"},{"id":97745411,"identity":"7aece4ed-9a6e-44e4-bfc3-b734d9d3c6ee","added_by":"auto","created_at":"2025-12-09 00:24:37","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":14561,"visible":true,"origin":"","legend":"","description":"","filename":"GraphicalAbstract.docx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/d1fb14d01e964e822f2325c1.docx"},{"id":97896325,"identity":"2092a51f-624a-46d9-ac3b-93c3e8c21bce","added_by":"auto","created_at":"2025-12-10 15:36:21","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":44873,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-7875969/v1/69e7956c30795564b698745d.docx"}],"financialInterests":"","formattedTitle":"Risk prediction models for hepatic encephalopathy in patients with liver cirrhosis: A systematic review and meta-analysis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHepatic encephalopathy (HE) is a neuropsychiatric complication arising from various acute and chronic liver diseases or abnormalities of portal-corporeal circulation shunting[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. In terms of its severity, the spectrum of HE ranges from subclinical changes in cognitive function, known as covert or minimal HE (CHE or MHE), to overt HE (OHE) with disorientation, confusion, and coma[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. It is estimated that 30% \u0026minus;\u0026thinsp;40% of patients with cirrhosis will experience at least one episode of OHE during their lifetime, and the recurrence rate following an initial episode remains high[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. The burdens and costs associated with HE for patients and the healthcare system are extensive and increasing[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eNumerous risk factors contribute to the development of HE. These factors include the severity of liver dysfunction (often assessed by Child-Pugh or model for end-stage liver disease (MELD) scores), a history of prior HE episodes, the presence of a transjugular intrahepatic portosystemic shunt (TIPS), gastrointestinal bleeding, infections, electrolyte disturbances (such as hyponatremia and hypokalemia), constipation, advanced age, and the use of specific medications like sedatives or diuretics[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. In addition, recent studies have emphasized the predictive value of factors such as sarcopenia, genetic factors, and specific laboratory indicators[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eVarious prediction models for HE in cirrhosis patients have been developed and evaluated in response to the critical need for better risk stratification. These models range from traditional scoring systems based on readily available clinical and laboratory parameters to more sophisticated approaches incorporating psychometric tests, radionics assessments, and advanced statistical techniques, including machine learning and artificial intelligence[\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Although there are increasing risk prediction models for HE in patients with cirrhosis, the quality and applicability of these models have not been formally and systematically reviewed. Therefore, this study aims to identify, appraise, and synthesize the evidence on the performance and predictors of existing prediction models for the risk of developing HE in patients with cirrhosis. The findings will provide valuable references for clinical practice and study design for future HE prediction.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThe protocol for this systematic review and meta-analysis was registered on PROSPERO (CRD420251040913).\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eSearch strategy\u003c/h2\u003e\u003cp\u003eTo conduct a comprehensive search, both Chinese and English databases were considered. The electronic databases searched included China National Knowledge Infrastructure (CNKI), WanFang Database, China Science and Technology Journal Database (VIP), SinoMed, PubMed, Embase, Web of Science, Scopus, and the Cochrane Library, from their establishment until April 12, 2025. Databases were searched using Medical Subject Headings (MeSH) and entry terms. No further search constraints were implemented. The search terms included \u0026ldquo;Liver Cirrhosis\u0026rdquo;, \u0026ldquo;Cirrhosis\u0026rdquo;, \u0026ldquo;Liver Fibrosis\u0026rdquo;, \u0026ldquo;Hepatic Encephalopathy\u0026rdquo;, \u0026ldquo;Hepatic Coma\u0026rdquo;, \u0026ldquo;Risk prediction model\u0026rdquo;, \u0026ldquo;Risk factor\u0026rdquo;, \u0026ldquo;Predictor\u0026rdquo;, \u0026ldquo;Model\u0026rdquo;, \u0026ldquo;Risk score\u0026rdquo;, and \u0026ldquo;Prediction tool\u0026rdquo;. The specific search strategy is shown in the Supplemental Materials. Furthermore, we manually searched references in the pertinent literature for additional potentially suitable studies.\u003c/p\u003e\u003cp\u003eWe employed the PICOTS framework for the systematic review, recommended by the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. This system helps frame the review's aim, search strategy, and study inclusion and exclusion criteria. The key items of the systematic review are described below:\u003c/p\u003e\u003cp\u003eP (Population): Patients with cirrhosis (whether or not TIPS treatment is accepted).\u003c/p\u003e\u003cp\u003eI (Intervention): Any prognostic model developed and published to predict HE risk in patients with cirrhosis.\u003c/p\u003e\u003cp\u003eC (Comparator): No competing model.\u003c/p\u003e\u003cp\u003eO (Outcome): The outcome focused on the occurrence of HE rather than other liver-related events and survival rate.\u003c/p\u003e\u003cp\u003eT (Timing): The outcome was predicted after evaluating basic information, laboratory indicators, genetic and imaging characteristics, and other factors without imposing any specific restrictions on the scope of the prediction.\u003c/p\u003e\u003cp\u003eS (Setting): The intended role of the risk prediction model is to individualize the prediction of the probability of developing HE in patients with cirrhosis to guide risk mitigation strategies.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eInclusion and exclusion criteria\u003c/h3\u003e\n\u003cp\u003eThe inclusion criteria for studies were: (1) patients aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years who met the liver cirrhosis diagnostic criteria. (2) the types of studies included case-control studies, cohort studies, and cross-sectional studies. (3) a prognostic prediction model was constructed. (4) the outcome of interest was HE, including the minimal, covert, and overt types. (5) articles written in English or Chinese.\u003c/p\u003e\u003cp\u003eThe exclusion criteria for studies were: (1) focusing only on risk factors without constructing a prediction model. (2) diagnostic model or non-original model. (3) HE data not reported separately. (4) the full text of the literature is not available or unable to extract data. (5) review, meta-analysis, case report, conference paper, dissertation, and other types of literature.\u003c/p\u003e\n\u003ch3\u003eLiterature selection and screening\u003c/h3\u003e\n\u003cp\u003eTwo researchers (ZX and XL) independently conducted the literature selection process. Initially, duplicate literature was identified utilizing EndNote 20 software and removed manually. Subsequently, the remaining studies were further evaluated according to their titles and abstracts to determine eligibility. Ultimately, following the application of the inclusion and exclusion criteria, the full text was reviewed to determine the studies included. Disagreements that emerged throughout the screening process were resolved through consultation and discussion with a third researcher (FZ).\u003c/p\u003e\n\u003ch3\u003eData extraction\u003c/h3\u003e\n\u003cp\u003eTwo researchers (ZX and XL) independently extracted the data, and any disagreements were resolved through consultation and discussion with a third researcher (FZ).\u003c/p\u003e\u003cp\u003eThe information extracted from all eligible articles was categorized into three groups: (1) Basic characteristics of the study: include the first author, publication year, cohort source, study design, main outcome, sample size, incidence of HE, age, gender distribution, TIPS status, research period, etiology, and patient type. (2) Basic information of the models: include the variable selection method, model development method, calibration method, missing data handling, validation method, final predictors, model presentation, and model performance. Potential indicators for assessing model calibration included the Brier score, calibration plot, calibration curve, and the Hosmer-Lemeshow test. We extracted data for each model separately if multiple models were described in the same article. (3) Distribution of predictors in the model: include the general information, laboratory indicators, comorbidities, imaging features, genetic features, Chinese medicine symptoms, and psychological tests.\u003c/p\u003e\n\u003ch3\u003eRisk of bias assessment\u003c/h3\u003e\n\u003cp\u003eTo evaluate the risk of bias (ROB) and applicability of the included studies, we employed the available version of the Prediction Model Risk of Bias Assessment Tool (PROBAST)[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. PROBAST enables a focused and transparent approach to assessing the ROB and applicability of studies that develop, validate, or update prediction models for individualized predictions[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. PROBAST is organized into the following four domains: participants, predictors, outcome, and analysis, which contain a total of 20 signaling questions. Each signaling question can be answered as \u0026ldquo;yes/probably yes\u0026rdquo;, \u0026ldquo;no/probably no\u0026rdquo;, or \u0026ldquo;no information\u0026rdquo;. If at least one signaling question in a domain is answered with \u0026ldquo;no/probably no\u0026rdquo;, that domain should be considered at high risk of bias. Overall bias can only be deemed low risk when all domains are evaluated as low risk of bias. Two researchers (ZX and XL) independently employed the PROBAST tool for assessment, and any disagreements were resolved through consultation and discussion with a third researcher (FZ).\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eWe employed Stata software (version 17.0) to conduct a meta-analysis of the area under the curve (AUC) for the externally validated prediction models. For studies reporting the c-statistic without specifying AUC, we treated the c-statistic as equivalent to AUC when the prediction model was designed for binary outcomes[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Meta-analysis of predictors was performed using Review Manager 5.4 software, and the combined odds ratio (OR) value and its 95% confidence intervals (CI) were calculated using the OR value as the effect indicator. Due to the large variety of predictors and the few studies reporting OR values, a meta-analysis was performed on only the predictors with the top ten ranked probabilities. In principle, the predictors with a combined number of studies of two or more were included. The I\u003csup\u003e2\u003c/sup\u003e statistic and the Cochrane\u0026rsquo;s Q test were used to evaluate the extent of heterogeneity. I\u003csup\u003e2\u003c/sup\u003e values\u0026thinsp;\u0026le;\u0026thinsp;25% indicate low heterogeneity, 25%\u0026lt; I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026le;\u0026thinsp;50% indicate moderate heterogeneity, and I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;50% indicate high heterogeneity. If P\u0026thinsp;\u0026lt;\u0026thinsp;0.1 and I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;50%, the random effects model was utilized; otherwise, heterogeneity was deemed acceptable, and the fixed effects model was chosen. In addition, the AUC value measures the effect size: 0.5\u0026ndash;0.7 indicates poor predictive performance, 0.7\u0026ndash;0.9 indicates moderate predictive performance, and 0.9-1.0 indicates excellent predictive performance. Finally, Egger\u0026rsquo;s test was used to assess publication bias. The Egger\u0026rsquo;s test with a p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered publication bias.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003eLiterature selection\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA-2020) flowchart, illustrating the exhaustive search process and results.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eA total of 14,728 literature records were initially retrieved. After removing 5,889 duplicate records, 8,839 studies were screened for titles and abstracts. Following the preliminary screening, 94 articles were included for full-text assessment, resulting in the exclusion of 56 studies, one of which included a non-cirrhosis population, ten focused on incorrect outcomes, seven were not available in full text or unable to extract data, 27 reported diagnostic models or non-original models, and 11 republished the same content. The Supplementary Material describes information about studies excluded after assessing the full text. Ultimately, 38 best prediction models from 38 studies were included in this review[\u003cspan additionalcitationids=\"CR9 CR10 CR11\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan additionalcitationids=\"CR18 CR19 CR20 CR21 CR22 CR23 CR24 CR25 CR26 CR27 CR28 CR29 CR30 CR31 CR32 CR33 CR34 CR35 CR36 CR37 CR38 CR39 CR40 CR41 CR42 CR43 CR44 CR45 CR46 CR47 CR48\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eStudy characteristics\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;1 summarizes the specific characteristics of the 38 included studies. These studies were published between 2015 and 2025, of which 30 were conducted in China (13 studies published in Chinese[\u003cspan additionalcitationids=\"CR38 CR39 CR40 CR41 CR42 CR43 CR44 CR45 CR46 CR47 CR48\" citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]), four in America, three in Europe, and one in Japan. Of the included studies, six were prospective (including four multicenter studies), and 32 were retrospective, of which 26 were conducted in a single center. In terms of study outcomes, 17 studies focused on the development of OHE, four on the development of CHE and MHE, the remaining studies were unspecified, and 17 specifically addressed HE after TIPS. The etiology of cirrhosis in the studies was broad, including alcohol, viral, non-alcoholic fatty liver disease (NAFLD), autoimmune, cholestatic, and other, with three studies all focusing on patients with hepatitis B, and five studies did not specify the etiology of cirrhosis. The sample sizes of the included studies ranged from 108 to 1,979, and the incidence of HE ranged from 7.2% to 50.4%.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eModels information\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;2 presents detailed information about the models in the included studies. Among the included studies, the most common modeling method used was logistic regression analysis (26 studies), followed by machine learning (8 studies) and Cox regression analysis (5 studies). Among the machine learning methods, the most common was support vector machine (SVM, seven studies), and the rest included random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), decision tree, and artificial neural network (ANN). In addition, nine studies used two or more modeling methods. Eighteen studies reported calibration methods (including calibration plots, calibration curves, Brier score, and the Hosmer-Lemeshow test), and five studies reported the handling of missing data. Among the included studies, the majority of the models were internally or externally validated. Of these, 25 studies were internally validated, 15 were externally validated, and 10 were both internally and externally validated, while eight studies did not undergo any validation after development. The primary method of model presentation involved nomograms, including 14 studies. The reported AUC or c-statistic values in the models ranged from 0.667 to 0.969.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003ePredictors in the model\u003c/h2\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the distribution of predictors in the models. Among the included models, the most commonly used predictors were age and Child-Pugh grade/score, which appeared in 16 and 12 models, respectively. Other commonly used predictors included imaging characteristics, albumin, total bilirubin, previous HE, and creatinine. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e illustrates at least 10 predictors by predictor category.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eRisk of bias assessment\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;3 and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e summarize the risk of bias and applicability of the included studies. Thirty-four studies were found to have a high risk of bias, and 27 studies raised applicability concerns.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eIn the participant domain, four studies were considered to have a high risk of bias, primarily due to the use of inappropriate data sources, with the included studies being retrospective non-cohort studies[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. In the predictor domain, eight studies were identified as having a high risk of bias due to potential unblinding in the assessment of predictors[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. In the outcome domain, six studies were found to be at a high risk of bias, while eleven studies were classified as having an unclear risk of bias due to the lack of blinded assessments between outcome and predictor or potential unblinding[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan additionalcitationids=\"CR25 CR26\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan additionalcitationids=\"CR30\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan additionalcitationids=\"CR37\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn the analysis domain, the majority of studies containing 31 items were found to be at high risk of bias, and three studies had an uncertain risk of bias. Of them, four studies had insufficient sample sizes of less than 20 events per variable (EPV)[\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]; three did not include all participants[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]; twenty-two did not avoid selecting predictors based on univariate analysis[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan additionalcitationids=\"CR40 CR41 CR42 CR43 CR44 CR45 CR46 CR47 CR48\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]; eighteen did not account for model overfitting and optimism in model performance[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan additionalcitationids=\"CR25 CR26\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan additionalcitationids=\"CR36\" citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan additionalcitationids=\"CR43 CR44\" citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn terms of applicability assessment, 27 studies were classified as high risk, and 11 studies were classified as low risk. In the participant domain, 26 studies were deemed to have a high applicability risk, as some studies solely included participants with TIPS or non-TIPS. In the predictor domain, three studies were considered high risk for applicability because there were concerns about predictor assessment[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. In the outcome domain, high concerns about applicability existed for three studies because one study focused only on advanced HE and two studies focused only on HE within three months[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eMeta-analysis\u003c/h2\u003e\u003cp\u003eAmong the studies included, only eight reported the model\u0026rsquo;s performance (AUC or c-statistic) along with its 95% CI after external validation. Of them, two studies reported the c-statistic based on survival analysis models[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], four studies developed models using logistic regression and reported AUC[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e], one study used a machine learning model[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e], and one study used a Fine-Gray competing risk model[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Consequently, only four studies that also employed logistic regression models were incorporated into the meta-analysis. The combined AUC value based on the fixed-effects model was 0.802 (95% CI: 0.785\u0026ndash;0.820) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). The I\u003csup\u003e2\u003c/sup\u003e value was 0% (P\u0026thinsp;=\u0026thinsp;0.658), indicating low heterogeneity between studies, and the P-value for Egger\u0026rsquo;s test was 0.193, suggesting no significant publication bias.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eOf at least 70 predictors, only 10 included\u0026thinsp;\u0026ge;\u0026thinsp;5 studies and were thus classified as major predictors for inclusion in the meta-analysis. However, a meta-analysis of the MELD score and creatinine could not be performed due to the number of studies reporting OR values being fewer than two. The meta-analysis showed that age (OR\u0026thinsp;=\u0026thinsp;1.04, 95% CI: 1.03, 1.05), prior HE (OR\u0026thinsp;=\u0026thinsp;4.42, 95% CI: 2.67, 7.31), low albumin (OR\u0026thinsp;=\u0026thinsp;1.78, 95% CI: 1.25, 2.56), total bilirubin (OR\u0026thinsp;=\u0026thinsp;2.22, 95% CI: 1.73, 2.85), Child-Pugh grade/score (OR\u0026thinsp;=\u0026thinsp;2.41, 95% CI: 1.87, 3.09; OR\u0026thinsp;=\u0026thinsp;1.65, 95% CI: 1.16, 2.33, respectively), ascites (OR\u0026thinsp;=\u0026thinsp;1.96, 95% CI: 1.48, 2.60), and co-infection (OR\u0026thinsp;=\u0026thinsp;2.57, 95% CI: 1.66, 3.98) were significant predictors of HE in cirrhosis (P\u0026thinsp;\u0026lt;\u0026thinsp;0.01). The OR for serum sodium was 0.92 (95% CI: 0.80\u0026ndash;1.05) (P\u0026thinsp;=\u0026thinsp;0.21), indicating a not statistically significant difference. However, the analysis included only two studies, and the heterogeneity was high (I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;67%, P\u0026thinsp;=\u0026thinsp;0.08), which requires a cautious interpretation of the results. In addition, the heterogeneity of Child-Pugh score was also high (I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;90%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The results of the Egger\u0026rsquo;s test indicated the presence of publication bias in studies about Child-Pugh grade and ascites, with P-values of 0.002 and 0.026, respectively. Table\u0026nbsp;4 and Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e present the results of the meta-analysis of risk factors for HE in patients with cirrhosis.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eHE is a common and serious complication in patients with cirrhosis, occurring in 30% to 45% of cases, and is an independent risk factor for death in cirrhotic patients[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Accurately predicting HE risk is crucial for guiding clinical decisions, optimizing resource allocation, and improving patient outcomes. Recently, prediction models based on clinical, radiomic, and psychological tests have emerged as key tools for the early identification of HE. This review aimed to comprehensively assess the performance of existing models for predicting HE risk in patients with cirrhosis and to summarize their value for clinical application, although most of these are based on Chinese patient data. This study included 38 studies, each reporting a model with the best predictive performance involving AUC values ranging from 0.667 to 0.969. However, according to the PROBAST checklist, most studies had a high risk of bias and applicability concerns, which may be related to the study design, patient population characteristics, definition of HE, and insufficient model validation. The combined AUC value of the models included in the meta-analysis was 0.802 (95% CI: 0.785\u0026ndash;0.820); however, the small number of studies involved may affect the extrapolation of the results. Furthermore, several studies failed to adequately adhere to the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) statement[\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. Specifically, the methods for variable selection, handling of missing data, calibration methods, model performance metrics, and confidence intervals were not sufficiently detailed, potentially introducing bias and uncertainty into the models.\u003c/p\u003e\u003cp\u003eLogistic regression is a widely used modeling method, with 26 studies in this review employing logistic regression analysis, followed by 8 studies utilizing machine learning. While machine learning models have the potential to enhance predictive power and accuracy over traditional logistic regression models[\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e], a systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models[\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. Additionally, machine learning models often lack suitable representation tools, which may hinder their further clinical application[\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Given the current evidence, it remains challenging to determine which model is superior for predicting the risk of HE when comparing prediction models developed using various methods. Therefore, in the future development and application of predictive models, it is crucial to thoroughly consider the model\u0026rsquo;s performance characteristics, the relevant population, and its practicality in clinical scenarios.\u003c/p\u003e\u003cp\u003eIn the prediction models reported in this review, common predictors included age, previous HE, albumin, total bilirubin (TBIL), Child-Pugh grade/score, MELD score, serum sodium, creatinine, ascites, and infection. These predictors are easily accessible and cost-effective, making them valuable for future research and clinical practice. Firstly, age serves as an independent risk factor for HE in individuals with cirrhosis. Older patients often experience sarcopenia, which can reduce the ability of muscle mitochondria to detoxify ammonia, and ageing itself can negatively affect brain function, further increasing the risk of HE[\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. Secondly, patients with a history of OHE episodes have a 42% risk of recurrence within one year[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. The persistence of cognitive dysfunction and porto-systemic shunting diminishes the body\u0026rsquo;s capacity to compensate for these issues, highlighting the need for secondary prevention and health education for those with a previous HE history. Moreover, traditional scoring systems, such as Child-Pugh and MELD, are frequently employed to estimate the risk of HE, as they partially indicate the extent of liver failure. The Child-Pugh grading system considers the patient\u0026rsquo;s overall condition, ascites, TBIL, albumin levels, and prothrombin time. In contrast, the MELD score incorporates TBIL, creatinine levels, international normalized ratio (INR), and cirrhosis etiology to evaluate the liver functional reserve and prognosis for patients with chronic liver disease. Both scoring systems include TBIL, indicating its significance in predicting liver function. Research suggests that TBIL is an independent factor influencing mortality among HE patients[\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Elevated levels of bilirubin, particularly unconjugated bilirubin, can contribute to or worsen the development of HE through various mechanisms, including direct neurotoxicity, exacerbation of oxidative stress and inflammation, impairment of mitochondrial function, and effects on blood-brain barrier permeability[\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eAlbumin is often used to assess nutritional status, and patients with decompensated cirrhosis are usually associated with varying degrees of malnutrition and sarcopenia, which can contribute to the conversion of gluconeogenesis and increase ammonia production when the body is in a state of negative nitrogen balance[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. Studies have indicated that albumin is an independent risk factor for the development of HE and that albumin infusion reduces the incidence and severity of HE in patients with cirrhosis[\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. In addition, a prospective study showed that the presence of hyponatremia is a major risk factor for the development of OHE in cirrhotic patients[\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. The study suggests that developing hyponatremia in the context of disturbed osmotic homeostasis in the cirrhotic brain may represent a second osmotic blow to astrocytes[\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. The OR value for serum sodium combined in this meta-analysis was 0.92 (95% CI: 0.80\u0026ndash;1.05), a difference that was not statistically significant, which is not in line with previous studies, mainly due to biased results from the small number of included studies. Elevated serum creatinine levels usually reflect impaired renal function, which exacerbates the accumulation of toxins in the body and systemic inflammation, thereby contributes to the development of HE. Including creatinine in the MELD score also signals the important role of renal factors in the prognosis of liver cirrhosis. Similarly, the development of ascites in cirrhotic patients is closely associated with hepatic decompensation and exacerbation of portal hypertension, which is one of the key drivers of hepatic encephalopathy[\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e]. Finally, infection is a recognized HE trigger in cirrhotic patients, most commonly seen in spontaneous bacterial peritonitis, associated with hyperammonemia due to intestinal bacterial overgrowth[\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e]. At the same time, infections can further worsen astrocyte dysfunction already caused by hyperammonemia through the direct action of inflammatory mediators, cerebral microvascularisation and endothelial metabolism[\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn addition to traditional clinical and laboratory indicators, emerging models have started to incorporate radiomic features, genetic factors, and psychological assessments into their prediction systems, enhancing their ability to predict the occurrence of HE. Among the studies examined, the radiomic factors included computed tomography (CT) features related to visceral adipose tissue, the liver, the spleen, and the liver-related vascular morphology[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. However, it is important to note that obtaining these factors through imaging techniques can be cumbersome and costly, which means their use in clinical practice requires further evaluation. Genetic risk factors for HE have been previously identified, and cirrhotic patients carrying GLS (encoding glutaminase) variants will have a significantly increased incidence of HE[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Integrating genetic background information into prediction models may prove beneficial for forecasting HE in cirrhotic patients, but further validation in diverse populations with varying genetic backgrounds is essential[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. It is well established that the psychometric hepatic encephalopathy score (PHES) is the gold standard for diagnosing MHE[\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e]. Recent studies have confirmed that the PHES can predict the development of OHE and mortality in cirrhotic patients[\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e]. However, psychometric tests can be highly subjective, reliant on the patient, and time-consuming. Future research should explore how these tests can be simplified and integrated into predictive models.\u003c/p\u003e\u003cp\u003eOur study has several potential limitations. First, most of the included studies were conducted in China, which may limit the generalizability of the findings to other populations. This suggests that adjustments may be needed when applying these models in different regions, highlighting the need to develop risk prediction models in more diverse populations in the future. Second, only four studies that were externally validated were ultimately included in the meta-analysis due to insufficient information reported by some of the studies. This could lead to bias in assessing the heterogeneity between studies and may reduce the statistical efficacy related to publication bias. Additionally, there may be differences in the objectivity and consistency of diagnosing or staging HE across studies, which can impact the pooled results. However, these issues do not affect the systematic evaluation of prediction models. Finally, we were unable to conduct subgroup analyses for all potential confounders due to limitations in data availability.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis systematic review includes thirty-eight studies and models. The four externally validated models achieved a combined AUC of 0.802 (95% CI: 0.785\u0026ndash;0.820), indicating moderate predictive performance. This study showed that age, prior HE, low albumin, total bilirubin, Child-Pugh grade/score, ascites, and co-infection were significant predictors of HE in cirrhosis. However, according to the PROBAST checklist, thirty-four studies were found to have a high risk of bias, and twenty-seven studies raised concerns regarding their applicability. Therefore, the clinical utility of the model requires further validation. Future research should focus on developing new models through optimised study design and analysis, increased sample sizes and external validation, and applying them to clinical practice.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eHE Hepatic encephalopathy\u003c/p\u003e\u003cp\u003eCHE Covert hepatic encephalopathy\u003c/p\u003e\u003cp\u003eMHE Minimal hepatic encephalopathy\u003c/p\u003e\u003cp\u003eOHE Overt hepatic encephalopathy\u003c/p\u003e\u003cp\u003eMELD Model for end-stage liver disease\u003c/p\u003e\u003cp\u003eTIPS Transjugular intrahepatic portosystemic shunt\u003c/p\u003e\u003cp\u003eCNKI China National Knowledge Infrastructure\u003c/p\u003e\u003cp\u003eVIP China Science and Technology Journal Database\u003c/p\u003e\u003cp\u003eMeSH Medical Subject Headings\u003c/p\u003e\u003cp\u003eCHARMS Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies\u003c/p\u003e\u003cp\u003eROB Risk of bias\u003c/p\u003e\u003cp\u003ePROBAST Prediction Model Risk of Bias Assessment Tool\u003c/p\u003e\u003cp\u003eAUC Area under the curve\u003c/p\u003e\u003cp\u003eOR Odds ratio\u003c/p\u003e\u003cp\u003eCI Confidence intervals\u003c/p\u003e\u003cp\u003ePRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses\u003c/p\u003e\u003cp\u003eNAFLD Non-alcoholic fatty liver disease\u003c/p\u003e\u003cp\u003eSVM Support vector machine\u003c/p\u003e\u003cp\u003eRF Random forest\u003c/p\u003e\u003cp\u003eXGBoost Extreme gradient boosting\u003c/p\u003e\u003cp\u003eLightGBM Light gradient boosting machine\u003c/p\u003e\u003cp\u003eANN Artificial neural network\u003c/p\u003e\u003cp\u003eTRIPOD Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis\u003c/p\u003e\u003cp\u003eTBIL Total bilirubin\u003c/p\u003e\u003cp\u003eINR International normalized ratio\u003c/p\u003e\u003cp\u003eCT Computed tomography\u003c/p\u003e\u003cp\u003ePHES Psychometric hepatic encephalopathy score\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analysed during this study are included in this published article and its supplementary information files.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by National Natural Science Foundation of China (grant number 82070569).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZhi-Le Xiao: Conceptualization and methodology, Data curation, Formal analysis, Visualization and validation, Original draft writing. Xiao-Nan Li: Data curation, Formal analysis, Visualization and validation. Feng Zhou: Conceptualization and methodology, Data curation, Formal analysis, Visualization and validation, Project administration. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eH\u0026auml;ussinger D, Dhiman RK, Felipo V, et al. Hepatic encephalopathy. Nat Rev Dis Primer. 2022;8:43.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMontagnese S, Rautou P-E, Romero-G\u0026oacute;mez M, Larsen FS, Shawcross DL, Thabut D, Vilstrup H, Weissenborn K. EASL Clinical Practice Guidelines on the management of hepatic encephalopathy. J Hepatol. 2022;77:807\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLouissaint J, Deutsch-Link S, Tapper EB. Changing Epidemiology of Cirrhosis and Hepatic Encephalopathy. Clin Gastroenterol Hepatol. 2022;20:S1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVidal-Cevallos P, Ch\u0026aacute;vez-Tapia NC, Uribe M. Current approaches to hepatic encephalopathy. Ann Hepatol. 2022;27:100757.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRose CF, Amodio P, Bajaj JS, Dhiman RK, Montagnese S, Taylor-Robinson SD, Vilstrup H, Jalan R. Hepatic encephalopathy: Novel insights into classification, pathophysiology and therapy. J Hepatol. 2020;73:1526\u0026ndash;47.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNardelli S, Lattanzi B, Merli M, Farcomeni A, Gioia S, Ridola L, Riggio O. Muscle Alterations Are Associated With Minimal and Overt Hepatic Encephalopathy in Patients With Liver Cirrhosis. Hepatology. 2019;70:1704\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMayer LB, Krawczyk M, Gr\u0026uuml;nhage F, Lammert F, Stokes CS. A genetic variant in the promoter of phosphate-activated glutaminase is associated with hepatic encephalopathy. J Intern Med. 2015;278:313\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBajaj JS, Tandon P, O\u0026rsquo;Leary JG, et al. Admission Serum Metabolites and Thyroxine Predict Advanced Hepatic Encephalopathy in a Multicenter Inpatient Cirrhosis Cohort. Clin Gastroenterol Hepatol. 2023;21:1031\u0026ndash;e10403.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTapper EB, Parikh ND, Sengupta N, Mellinger J, Ratz D, Lok AS-F, Su GL. A risk score to predict the development of hepatic encephalopathy in a population‐based cohort of patients with cirrhosis. Hepatology. 2018;68:1498\u0026ndash;507.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYi C, Wei Y, Zhengcui H, Ying X, Xiaoying L. Construction of an early warning model of hepatic encephalopathy in patients with liver cirrhosis based on psychological testing. J Clin Nurs. 2023;32:191\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCheng S, Hu G, Jin Z, Wang Z, Xue H. Prediction of Hepatic Encephalopathy After Transjugular Intrahepatic Portosystemic Shunt Based on CT Radiomic Features of Visceral Adipose Tissue. Acad Radiol. 2024;31:1849\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang H, Li X, Cao H, Cui Y, Luo Y, Liu J, Zhang Y. Using machine learning methods to predict hepatic encephalopathy in cirrhotic patients with unbalanced data. Comput Methods Programs Biomed. 2021;211:106420.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoons KGM, de Groot JAH, Bouwmeester W, Vergouwe Y, Mallett S, Altman DG, Reitsma JB, Collins GS. Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies: The CHARMS Checklist. PLoS Med. 2014;11:e1001744.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWolff RF, Moons KGM, Riley RD, Whiting PF, Westwood M, Collins GS, Reitsma JB, Kleijnen J, Mallett S. PROBAST: A Tool to Assess the Risk of Bias and Applicability of Prediction Model Studies. Ann Intern Med. 2019;170:51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMoons KGM, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, Reitsma JB, Kleijnen J, Mallett S. PROBAST: A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies: Explanation and Elaboration. Ann Intern Med. 2019;170:W1.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVerbakel JY, Steyerberg EW, Uno H, De Cock B, Wynants L, Collins GS, Van Calster B. ROC curves for clinical prediction models part 1. ROC plots showed no added value above the AUC when evaluating the performance of clinical prediction models. J Clin Epidemiol. 2020;126:207\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBallester MP, Tranah TH, Balcar L, et al. Development and validation of the AMMON-OHE model to predict risk of overt hepatic encephalopathy occurrence in outpatients with cirrhosis. J Hepatol. 2023;79:967\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCao J, Yang J, Ming Z, et al. A radiomics model of liver CT to predict risk of hepatic encephalopathy secondary to hepatitis B related cirrhosis. Eur J Radiol. 2020;130:109201.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen X, Huang M, Yu X, et al. Hepatic-associated vascular morphological assessment to predict overt hepatic encephalopathy before TIPS: a multicenter study. Hepatol Int. 2024;18:1238\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen X, Wang T, Ji Z, et al. 3D automatic liver and spleen assessment in predicting overt hepatic encephalopathy before TIPS: a multi-center study. Hepatol Int. 2023;17:1545\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGil-G\u0026oacute;mez A, Ampuero J, Rojas \u0026Aacute;, et al. Development and Validation of a Clinical-Genetic Risk Score to Predict Hepatic Encephalopathy in Patients With Liver Cirrhosis. Am J Gastroenterol. 2021;116:1238\u0026ndash;47.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eİnce O, \u0026Ouml;nder H, Gen\u0026ccedil;t\u0026uuml;rk M, Golzarian J, Young S. Machine Learning Insights: Predicting Hepatic Encephalopathy After TIPS Placement. Cardiovasc Intervent Radiol. 2023;46:1715\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi K, Cheng Y, Zhao R, Jiang H, Zhang L, Tong Y, Li S. Prediction of mortality and overt hepatic encephalopathy undergoing transjugular intrahepatic portosystemic shunt: a retrospective cohort study. Abdom Radiol. 2024;49:908\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi Y, He X, Pang H. A model to predict early hepatic encephalopathy in patients undergoing transjugular intrahepatic portosystemic shunt. Turk J Gastroenterol. 2019;30:702\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu J, Zhou C, Wang Y, Yang C, Shi Q, Huang S, Chen Y, Li T, Xiong B. The combination of Child\u0026ndash;Pugh score and quantitative CT-based spleen volume could predict the risk of hepatic encephalopathy after transjugular intrahepatic portosystemic shunt creation. Abdom Radiol. 2021;46:3464\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMiwa T, Hanai T, Nishimura K, Maeda T, Tajirika S, Imai K, Suetsugu A, Takai K, Yamamoto M, Shimizu M. A simple covert hepatic encephalopathy screening model based on blood biochemical parameters in patients with cirrhosis. PLoS ONE. 2022;17:e0277829.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRiggio O, Amodio P, Farcomeni A, et al. A Model for Predicting Development of Overt Hepatic Encephalopathy in Patients With Cirrhosis. Clin Gastroenterol Hepatol. 2015;13:1346\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTapper EB, Zhao L, Nikirk S, Baki J, Parikh ND, Lok AS, Waljee AK. Incidence and Bedside Predictors of the First Episode of Overt Hepatic Encephalopathy in Patients With Cirrhosis. Am J Gastroenterol. 2020;115:2017\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTong H, Gan C, Wei B, et al. Risk factors for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt creation in patients with liver cirrhosis. J Dig Dis. 2021;22:31\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang C, Teng Y, Gao J, Zhang Z, Li Y. Low adipose tissue index as an indicator of hepatic encephalopathy in cirrhotic patients following transjugular intrahepatic portosystemic shunt. Abdom Radiol. 2023;48:1454\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang C, Zhu X, Liu J, et al. Development and Validation of Prognostic Model to Estimate the Risk of Overt Hepatic Encephalopathy After TIPS Creation: A Multicenter Study. Clin Transl Gastroenterol. 2022;13:e00461.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang Y, Fu S, Cao B, et al. Prediction of overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt treatment: a cohort study. Hepatol Int. 2021;15:730\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYang Y, Liang X, Yang S, et al. Preoperative prediction of overt hepatic encephalopathy caused by transjugular intrahepatic portosystemic shunt. Eur J Radiol. 2022;154:110384.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYao C, Huang L, Wang M, et al. Establishment and validation of a nomogram model for riskprediction of hepatic encephalopathy: a retrospective analysis. Sci Rep. 2023;13:19544.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZeng X, Yin C, Sun CY, et al. Prevalence and risk factors of covert hepatic encephalopathy in cirrhotic patients: A multicenter study in China. J Dig Dis. 2023;24:122\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhong B-Y, Wang W-S, Shen J, Du H, Zhang S, Li W-C, Yin Y, Yang J, Ni C-F, Zhu X-L. Single-Centre Retrospective Training Cohort Using Artificial Intelligence for Prognostic Prediction of Encephalopathy, Mortality, and Liver Dysfunction after Early TIPS Creation. Cardiovasc Intervent Radiol. 2021;44:1597\u0026ndash;608.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen L, Pan C, Liu M, Yi L. Construction of risk assessment model for hepatic encephalopathy in patients with liver cirrhosis and its predictive value. China Mod Dr. 2023;61:29\u0026ndash;3273.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFu S, He B, Xu J. Establishment and validation of prediction model for cirrhosis-related hepatic encephalopathy by machine learning algorithm. Chin J Lab Med. 2025;48:93\u0026ndash;102.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGao X, Peng F, Li P. Analysis of risk factors of liver cirrhosis complicated with mild hepatic encephalopathy and construction of nomogram prediction model. Chin J Integr Tradit West Med Liver Dis. 2024;34:9\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang T, Wang G, Huang G, Zhang C. Prediction of different clinical outcomes in patients with cirrhosis after TIPS based on machine learning algorithm. Chin J Dig Med Imageology Electron Ed. 2022;12:4\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi XQ, Li Y, Ni YQ, Cao W, Yin TT, Lu R. A nomogram prediction model for individualized prediction of the risk of covert (minimal) hepatic encephalopathy occurrence in patients with liver cirrhosis. Zhonghua Gan Zang Bing Za Zhi. 2024;32:828\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi X, Li J, Sun W. Establishment and evaluation of a nomogram prediction model for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt in patients with liver cirrhosis and esophagogastric variceal bleeding. J Clin Hepatol. 2024;40:1605\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTan JT, Xu XM, He YX, Chao T, Jun G, Liu YY, Xiang SS, Zhao WL. Construction of prediction model of cirrhosis-related hepatic encephalopathy based on machine learning algorithm. Med J Chin Peoples Lib Army. 2021;46:354\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang L, Qin J, Yao X, Qi Q, Liu L, Tang S. Construction of a risk prediction model for overt hepatic encephalopathy after transjugular intrahepatic portosystemic shunt in patients with hepatitis B cirrhosis and portal hypertension. J Clin Hepatol. 2024;40:1149\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang N, Li J, Li X, Liang L, Gan X, Wang X. The development and application of a risk prediction model for hepatic encephalopathy in patients with liver cirrhosis. Chin J Nurs. 2019;54:805.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang X, Zhai M, Ren H, Li M, Quan D, Zhang Y, Liu J, Qiu L. An exploratory study of a classification model based on resampling and Voting heterogeneous integration in the prediction of the risk of liver cirrhosis complicated by hepatic encephalopathy. Chin J Health Stat. 2022;39:545\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWen J, Zhao Y, Wang S, et al. Construction and validation of a risk prediction model for liver cirrhosis complicated with minimal hepatic encephalopathy based on the involvement of traditional Chinese medicine factors. Chin J Liver Dis Electron Version. 2024;16:48\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYuan W, Li L, Han X, Liu J, Li B, Wu J. Clinical prediction modeling and efficacy analysis of the occurrence of hepatic encephalopathy after TIPS. Mod Dig Interv. 2021;26:874\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang J, Wang Z, Yang J, Wang X, Zhao Y. Construction and validation of the risk prediction model for secondary hepatic encephalopathy in patients with chronic hepatitis B and cirrhosis. Chin Hepatol. 2023;28:916\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eElwir S, Rahimi RS. Hepatic Encephalopathy: An Update on the Pathophysiology and Therapeutic Options. J Clin Transl Hepatol. 2017;5:142\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBajaj JS, O\u0026rsquo;Leary JG, Tandon P, et al. Hepatic Encephalopathy Is Associated With Mortality in Patients With Cirrhosis Independent of Other Extrahepatic Organ Failures. Clin Gastroenterol Hepatol. 2017;15:565\u0026ndash;e5744.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCollins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594\u0026ndash;7594.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRajkomar A, Dean J, Kohane I. Machine Learning in Medicine. N Engl J Med. 2019;380:1347\u0026ndash;58.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChurpek MM, Yuen TC, Winslow C, Meltzer DO, Kattan MW, Edelson DP. Multicenter Comparison of Machine Learning Methods and Conventional Regression for Predicting Clinical Deterioration on the Wards. Crit Care Med. 2016;44:368\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChristodoulou E, Ma J, Collins GS, Steyerberg EW, Verbakel JY, Van Calster B. A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models. J Clin Epidemiol. 2019;110:12\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFu H, Hou D, Xu R, You Q, Li H, Yang Q, Wang H, Gao J, Bai D. Risk prediction models for deep venous thrombosis in patients with acute stroke: A systematic review and meta-analysis. Int J Nurs Stud. 2024;149:104623.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWelch N, Attaway A, Bellar A, Alkhafaji H, Vural A, Dasarathy S. Compound Sarcopenia in Hospitalized Patients with Cirrhosis Worsens Outcomes with Increasing Age. Nutrients. 2021;13:659.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu X, Ding H, Li W, et al. Chinese Guidelines on the Management of Hepatic Encephalopathy in Cirrhosis (2024). J Clin Transl Hepatol. 2025;000:000\u0026ndash;000.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCui Y, Guan S, Ding J, He Y, Li Q, Wang S, Sun H. Establishment and evaluation of a model for predicting 3-month mortality in Chinese patients with hepatic encephalopathy. Metab Brain Dis. 2019;34:213\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePranty AI, Shumka S, Adjaye J. Bilirubin-Induced Neurological Damage: Current and Emerging iPSC-Derived Brain Organoid Models. Cells. 2022;11:2647.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmodio P, Bemeur C, Butterworth R, Cordoba J, Kato A, Montagnese S, Uribe M, Vilstrup H, Morgan MY. The nutritional management of hepatic encephalopathy in patients with cirrhosis: International society for hepatic encephalopathy and nitrogen metabolism consensus. Hepatology. 2013;58:325\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBai Z, Guo X, Tacke F, Li Y, Li H, Qi X. Association of serum albumin level with incidence and mortality of overt hepatic encephalopathy in cirrhosis during hospitalization. Ther Adv Gastroenterol. 2019;12:1756284819881302.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBai Z, Bernardi M, Yoshida EM, Li H, Guo X, M\u0026eacute;ndez-S\u0026aacute;nchez N, Li Y, Wang R, Deng J, Qi X. Albumin infusion may decrease the incidence and severity of overt hepatic encephalopathy in liver cirrhosis. Aging. 2019;11:8502\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuevara M, Baccaro ME, Torre A, et al. Hyponatremia Is a Risk Factor of Hepatic Encephalopathy in Patients With Cirrhosis: A Prospective Study With Time-Dependent Analysis. Am J Gastroenterol. 2009;104:1382\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eThomsen KL, S\u0026oslash;rensen M, Kj\u0026aelig;rgaard K, Eriksen PL, Lauridsen MM, Vilstrup H. Cerebral Aspects of Portal Hypertension. Clin Liver Dis. 2024;28:541\u0026ndash;54.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePiotrowski D, Boroń-Kaczmarska A. Bacterial infections and hepatic encephalopathy in liver cirrhosis\u0026ndash;prophylaxis and treatment. Adv Med Sci. 2017;62:345\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuimar\u0026atilde;es L, Piedade J, Duarte J, Baldin C, Victor L, Costa B, Veiga Z, Alc\u0026acirc;ntara C, Fernandes F, Pereira G. Hepatic Encephalopathy in Cirrhotic Patients With Bacterial Infections: Frequency, Clinical Characteristics, and Prognostic Relevance. J Clin Exp Hepatol. 2023;13:559\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJaffe A, Lim JK, Jakab SS. Pathophysiology of Hepatic Encephalopathy. Clin Liver Dis. 2020;24:175\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmerican Association for the Study of Liver Diseases, European Association for the Study of the Liver. Hepatic Encephalopathy in Chronic Liver Disease: 2014 Practice Guideline by the European Association for the Study of the Liver and the American Association for the Study of Liver Diseases. J Hepatol. 2014;61:642\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLabenz C, Toenges G, Schattenberg JM, Nagel M, Huber Y, Marquardt JU, Labenz J, Galle PR, W\u0026ouml;rns M-A. Outcome Prediction of Covert Hepatic Encephalopathy in Liver Cirrhosis: Comparison of Four Testing Strategies. Clin Transl Gastroenterol. 2020;11:e00172.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 4 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"liver cirrhosis, hepatic encephalopathy, risk prediction model, systematic review, meta-analysis","lastPublishedDoi":"10.21203/rs.3.rs-7875969/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7875969/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eTo conduct a systematic review and meta-analysis of existing prediction models for hepatic encephalopathy (HE) in patients with cirrhosis.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eChina National Knowledge Infrastructure (CNKI), WanFang Database, China Science and Technology Journal Database (VIP), SinoMed, PubMed, Embase, Web of Science, Scopus, and the Cochrane Library databases were searched for studies on prediction models for the risk of HE in cirrhosis from inception to April 12, 2025. Two researchers independently conducted the literature search and data extraction, and the quality of the literature was evaluated using the Prediction Model Risk of Bias Assessment Tool (PROBAST)\u0026mdash;meta-analysis using Review Manager 5.4 and Stata 17.0 software.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eThirty-eight best prediction models from thirty-eight studies were ultimately included in this review. Among them, 17 studies predicted HE after transjugular intrahepatic portosystemic shunt (TIPS). The incidence of HE ranged from 7.2% to 50.4%. The most commonly used predictors were age and Child-Pugh grade/score. The reported area under the curve (AUC) or c-statistic values ranged from 0.667 to 0.969. Thirty-four studies were found to have a high risk of bias, and 27 studies raised applicability concerns, primarily due to inappropriate data sources, limitations in the domain of analysis, and homogenous study populations. Four externally validated logistic regression models had a combined AUC of 0.802 (95% CI: 0.785\u0026ndash;0.820), indicating moderate predictive performance. In meta-analysis, age (OR\u0026thinsp;=\u0026thinsp;1.04, 95% CI: 1.03, 1.05), prior HE (OR\u0026thinsp;=\u0026thinsp;4.42, 95% CI: 2.67, 7.31), low albumin (OR\u0026thinsp;=\u0026thinsp;1.78, 95% CI: 1.25, 2.56), total bilirubin (OR\u0026thinsp;=\u0026thinsp;2.22, 95% CI: 1.73, 2.85), Child-Pugh grade/score (OR\u0026thinsp;=\u0026thinsp;2.41, 95% CI: 1.87, 3.09; OR\u0026thinsp;=\u0026thinsp;1.65, 95% CI: 1.16, 2.33, respectively), ascites (OR\u0026thinsp;=\u0026thinsp;1.96, 95% CI: 1.48, 2.60), and co-infection (OR\u0026thinsp;=\u0026thinsp;2.57, 95% CI: 1.66, 3.98) were significant predictors of HE in cirrhosis (P\u0026thinsp;\u0026lt;\u0026thinsp;0.01).\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003ePrediction models for estimating the risk of incident HE with cirrhosis demonstrate moderate discrimination performance, while with a high overall risk of bias and a lack of clinical effectiveness research. Future research should focus on developing new models through optimised study design and analysis, increased sample sizes and external validation, and applying them to clinical practice.\u003c/p\u003e\u003ch2\u003eRegistration:\u003c/h2\u003e\u003cp\u003eThe protocol for this review was registered on PROSPERO (CRD420251040913).\u003c/p\u003e","manuscriptTitle":"Risk prediction models for hepatic encephalopathy in patients with liver cirrhosis: A systematic review and meta-analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-09 00:24:32","doi":"10.21203/rs.3.rs-7875969/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e6f9c852-7aff-41c6-9803-eb983068a397","owner":[],"postedDate":"December 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-02T12:14:13+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-09 00:24:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7875969","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7875969","identity":"rs-7875969","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}