Development of a diagnostic predictive model for determining child stunting in Malawi: a comparative analysis of variable selection approaches.

BACKGROUND: Childhood stunting is a major indicator of child malnutrition and a focus area of Global Nutrition Targets for 2025 and Sustainable Development Goals. Risk factors for childhood stunting are well studied and well known and could be used in a risk prediction model for assessing whether a child is stunted or not. However, the selection of child stunting predictor variables is a critical step in the development and performance of any such prediction model. This paper compares the performance of child stunting diagnostic predictive models based on predictor variables selected using a set of variable selection methods. METHODS: Firstly, we conducted a subjective review of the literature to identify determinants of child stunting in Sub-Saharan Africa. Secondly, a multivariate logistic regression model of child stunting was fitted using the identified predictors on stunting data among children aged 0-59 months in the Malawi Demographic Health Survey (MDHS 2015-16) data. Thirdly, several reduced multivariable logistic regression models were fitted depending on the predictor variables selected using seven variable selection algorithms, namely backward, forward, stepwise, random forest, Least Absolute Shrinkage and Selection Operator (LASSO), and judgmental. Lastly, for each reduced model, a diagnostic predictive model for the childhood stunting risk score, defined as the child propensity score based on derived coefficients, was calculated for each child. The prediction risk models were assessed using discrimination measures, including area under-receiver operator curve (AUROC), sensitivity and specificity. RESULTS: The review identified 68 predictor variables of child stunting, of which 27 were available in the MDHS 2016-16 data. The common risk factors selected by all the variable selection models include household wealth index, age of the child, household size, type of birth (singleton/multiple births), and birth weight. The best cut-off point on the child stunting risk prediction model was 0.37 based on risk factors determined by the judgmental variable selection method. The model's accuracy was estimated with an AUROC value of 64% (95% CI: 60%-67%) in the test data. For children residing in urban areas, the corresponding AUROC was AUC = 67% (95% CI: 58-76%), as opposed to those in rural areas, AUC = 63% (95% CI: 59-67%). CONCLUSION: The derived child stunting diagnostic prediction model could be useful as a first screening tool to identify children more likely to be stunted. The identified children could then receive necessary nutritional interventions.

Diagnostic accuracy of enhanced liver fibrosis test for nonalcoholic steatohepatitis-related fibrosis: Multicenter study.

AIM: The enhanced liver fibrosis (ELF) test is a noninvasive method for diagnosing hepatic fibrosis in patients with nonalcoholic fatty liver disease (NAFLD). This multicenter cohort study aimed to evaluate the accuracy of the ELF test and compare it with other noninvasive tests in Japan. METHODS: We analyzed 371 Japanese patients with biopsy-proven NAFLD. We constructed area under the receiver operator characteristic curves (AUROC) to determine the diagnostic accuracies of the ELF test, the Mac-2-binding protein glycosylation isomer (M2BPGi), the Fibrosis-4 (FIB-4) index, and combinations of these indices. RESULTS: In patients with F0/F1/F2/F3/F4 fibrosis, the median values of the ELF test were 8.98/9.56/10.39/10.92/11.41, respectively. The AUROCs of the ELF test for patients with F0 versus F1-4, F0-1 versus F2-4, F0-2 versus F3-4, and F0-3 versus F4 fibrosis were 0.825/0.817/0.802/0.812, respectively. The AUROCs of the ELF test were greater than those of the FIB-4 index and M2BPGi at each fibrosis stage. Respective low and high cut-off values yielded sensitivities and specificities for predicting advanced fibrosis (≥F3) of 91.1% and 50.8%, and 38.5% and 92.8%, respectively. For F3 or F4 fibrosis, the combined values from the ELF test and FIB-4 index showed a sensitivity of 98.5%, and the combined values from the ELF test and M2BPGi assay showed a specificity of 97.5%. CONCLUSIONS: In Japan, the ELF test predicts NAFLD-related fibrosis from its early stages. The diagnostic ability of the ELF test was not inferior to that of other indices, and the combined values of ELF plus other indices were more accurate.

Machine Learning applied to student attentiveness detection: Using emotional and non-emotional measures.

Electronic learning (e-learning) is considered the new norm of learning. One of the significant drawbacks of e-learning in comparison to the traditional classroom is that teachers cannot monitor the students' attentiveness. Previous literature used physical facial features or emotional states in detecting attentiveness. Other studies proposed combining physical and emotional facial features; however, a mixed model that only used a webcam was not tested. The study objective is to develop a machine learning (ML) model that automatically estimates students' attentiveness during e-learning classes using only a webcam. The model would help in evaluating teaching methods for e-learning. This study collected videos from seven students. The webcam of personal computers is used to obtain a video, from which we build a feature set that characterizes a student's physical and emotional state based on their face. This characterization includes eye aspect ratio (EAR), Yawn aspect ratio (YAR), head pose, and emotional states. A total of eleven variables are used in the training and validation of the model. ML algorithms are used to estimate individual students' attention levels. The ML models tested are decision trees, random forests, support vector machines (SVM), and extreme gradient boosting (XGBoost). Human observers' estimation of attention level is used as a reference. Our best attention classifier is the XGBoost, which achieved an average accuracy of 80.52%, with an AUROC OVR of 92.12%. The results indicate that a combination of emotional and non-emotional measures can generate a classifier with an accuracy comparable to other attentiveness studies. The study would also help assess the e-learning lectures through students' attentiveness. Hence will assist in developing the e-learning lectures by generating an attentiveness report for the tested lecture.

Development and internal validation of the HIV In-hospital Mortality Prediction (HIV-IMP) risk score.

BACKGROUND: Despite advances in availability and access to antiretroviral therapy (ART), HIV still ranks as a major cause of global mortality. Hence, the aim of this study was to develop and internally validate a risk score capable of accurately predicting in-hospital mortality in HIV-positive patients requiring hospital admission. METHODS: Consecutive HIV-positive patients presenting to the Charlotte Maxeke Johannesburg Academic Hospital adult emergency department between 7 July 2017 and 18 October 2018 were prospectively enrolled. Multivariate logistic regression was used to determine parameters for inclusion in the final risk score. Discrimination and calibration were assessed by means of the area under the receiver operating curve (AUROC) and the Hosmer-Lemeshow goodness-of-fit test, respectively. Internal validation was conducted using the regular bootstrap technique. RESULTS: The overall in-hospital mortality rate was 13.6% (n = 166). Eight predictors were included in the final risk score: ART non-adherence or not yet on ART, Glasgow Coma Scale < 15, respiratory rate > 20 breaths/min, oxygen saturation < 90%, white cell count < 4 × 109 /L, creatinine > 120 µmol/L, lactate > 2 mmol/L and albumin < 35 g/L. After internal validation, the risk score maintained good discrimination [AUROC 0.83, 95% confidence interval (CI): 0.78-0.88] and calibration (Hosmer-Lemeshow χ2 = 2.26, p = 0.895). CONCLUSION: The HIV In-hospital Mortality Prediction (HIV-IMP) risk score has overall good discrimination and calibration and is relatively easy to use. Further studies should be aimed at externally validating the score in varying clinical settings.

A note on estimating the Cox-Snell R2 from a reported C statistic (AUROC) to inform sample size calculations for developing a prediction model with a binary outcome.

In 2019 we published a pair of articles in Statistics in Medicine that describe how to calculate the minimum sample size for developing a multivariable prediction model with a continuous outcome, or with a binary or time-to-event outcome. As for any sample size calculation, the approach requires the user to specify anticipated values for key parameters. In particular, for a prediction model with a binary outcome, the outcome proportion and a conservative estimate for the overall fit of the developed model as measured by the Cox-Snell R2 (proportion of variance explained) must be specified. This proposal raises the question of how to identify a plausible value for R2 in advance of model development. Our articles suggest researchers should identify R2 from closely related models already published in their field. In this letter, we present details on how to derive R2 using the reported C statistic (AUROC) for such existing prediction models with a binary outcome. The C statistic is commonly reported, and so our approach allows researchers to obtain R2 for subsequent sample size calculations for new models. Stata and R code is provided, and a small simulation study.

Computational Prediction of Compound-Protein Interactions for Orphan Targets Using CGBVS.

A variety of Artificial Intelligence (AI)-based (Machine Learning) techniques have been developed with regard to in silico prediction of Compound-Protein interactions (CPI)-one of which is a technique we refer to as chemical genomics-based virtual screening (CGBVS). Prediction calculations done via pairwise kernel-based support vector machine (SVM) is the main feature of CGBVS which gives high prediction accuracy, with simple implementation and easy handling. We studied whether the CGBVS technique can identify ligands for targets without ligand information (orphan targets) using data from G protein-coupled receptor (GPCR) families. As the validation method, we tested whether the ligand prediction was correct for a virtual orphan GPCR in which all ligand information for one selected target was omitted from the training data. We have specifically expressed the results of this study as applicability index and developed a method to determine whether CGBVS can be used to predict GPCR ligands. Validation results showed that the prediction accuracy of each GPCR differed greatly, but models using Multiple Sequence Alignment (MSA) as the protein descriptor performed well in terms of overall prediction accuracy. We also discovered that the effect of the type compound descriptors on the prediction accuracy was less significant than that of the type of protein descriptors used. Furthermore, we found that the accuracy of the ligand prediction depends on the amount of ligand information with regard to GPCRs related to the target. Additionally, the prediction accuracy tends to be high if a large amount of ligand information for related proteins is used in the training.

Urinary Volatile Terpenes Analyzed by Gas Chromatography-Mass Spectrometry to Monitor Breast Cancer Treatment Efficacy in Mice.

Urinary volatile terpene (VT) levels are significantly altered with induced models of breast cancer in mice. The question arises whether VTs can detect the efficacy of antitumor treatments. BALB/c mice were injected with 4T1.2 murine tumor cells in the mammary pad or iliac artery to model localized breast cancer and induced bone metastasis. The effect of two dopaminergic antitumor agents was tested by conventional histology and altered VT levels. The headspace of urine specimens was analyzed by gas chromatography-mass spectrometry. In the localized model, the statistical significance (p < 0.05) was identified for 26% of VTs, and in the metastasis model, 19% of VTs. The authors discovered separate VT panels classifying localized/control [area under the curve (AUC) = 1.0] and metastasis/control (AUC = 0.98). Treatment samples were tested using these panels, which showed that mice treated with either agent were statistically significantly different from cancer samples, which is consistent with conventional analysis.

Peripheral quantitative computed tomography (pQCT)-based finite element analysis provides enhanced diagnostic performance in identifying non-vertebral fracture patients compared with dual-energy X-ray absorptiometry.

Due to limitations of the predominant clinical method for diagnosing osteoporosis, an engineering model based on a dedicated CT scanner for bone density and structure was applied in fracture patients and controls. Improved diagnostic performance was observed, which supports its potential use in future research and clinical practice. INTRODUCTION: Dual-energy X-ray absorptiometry (DXA), the predominant clinical method for diagnosing osteoporosis, has limitations in identifying individuals with increased fracture risk. Peripheral quantitative computed tomography (pQCT) provides additional information and can be used to generate finite element (FE) models from which bone strength properties can be estimated. We investigated the ability of pQCT-FE properties to distinguish peripheral low-trauma fracture patients from healthy controls, by comparison with DXA and standard pQCT. METHODS: One hundred and eight fracture patients (77 females aged 67.7 ± 7.9 years, 31 males aged 69.7 ± 8.9 years) were recruited from a hospital fracture liaison service. One hundred and twenty healthy community controls (85 females aged 69.8 ± 8.5 years, 35 males aged 68.9 ± 7.2 years) were recruited. RESULTS: Significant differences between groups were observed in pQCT-FE properties, especially at the 4% tibia site. Fracture odds increased most per standard deviation decrease in pQCT-FE at this location [shear stiffness estimate, kshear, in females, OR = 10.34, 95% CI (1.91, 43.98); bending stiffness estimate, kbend, in males, OR = 8.32, 95% CI (4.15, 33.84)]. Area under the receiver operating characteristics curve (AUROC) was observed to be highest with pQCT-FE properties at 4% the tibia site. In females, this was 0.83 for the pQCT-FE variable kshear, compared with 0.72 for DXA total hip bone density (TH aBMD) and 0.76 for pQCT tibia trabecular density (Trb vBMD); in males, this was 0.81 for the pQCT-FE variable kbend at the 4% tibia site, compared with 0.62 for TH aBMD and 0.71 for Trb vBMD. There were significant differences in AUROC between DXA and pQCT-FE variables in both females (p = 0.02) and males (p = 0.03), while no difference was observed in AUROC between primary pQCT and pQCT-FE variables. CONCLUSIONS: pQCT-FE modeling can provide enhanced diagnostic performance compared with DXA and, given its moderate cost, may be useful in clinical settings.

Levels of Cytokines in Serum Associate With Development of Hepatocellular Carcinoma in Patients With HCV Infection Treated With Direct-Acting Antivirals.

Concern has arisen about the development of hepatocellular carcinoma (HCC) in patients with hepatitis C virus (HCV) infection treated with direct-acting antivirals (DAAs). To identify patients at risk for HCC, we evaluated serum levels of immune mediators before, during, and after DAA treatment of HCV infection. Our study included 13 patients who developed HCC within 18 months after treatment (3 with HCC recurrence and 10 with new HCC) and 10 patients who did not develop HCC (controls), within at least 24 months of treatment (median, 26 months). We identified a set of 12 immune mediators (cytokines, growth factors, and apoptosis markers) whose levels were significantly higher in serum before DAA treatment of patients who eventually developed de novo HCC compared with controls. A panel of 9 cytokines, measured in serum before treatment (MIG, IL22, TRAIL, APRIL, VEGF, IL3, TWEAK, SCF, IL21), identified patients who developed de novo HCC with an area under the receiver operating characteristic curve value higher than 0.8. Further analyses of changes in levels of inflammatory cytokines during DAA treatment also provides important information about HCV-induced carcinogenesis and the effects of DAAs.

Molecular Ellipticity of Circulating Albumin-Bilirubin Complex Associates With Mortality in Patients With Severe Alcoholic Hepatitis.

BACKGROUND & AIMS: Hyperbilirubinemia and hypoalbuminemia are features of hepatic dysfunction that associate with disease severity. This is because hepatic insufficiency causes hypoalbuminemia, which indirectly increases the circulating levels of free bilirubin. Circular dichroism (CD) spectroscopy can be used to quantify the molecular ellipticity (ME) of the albumin-bilirubin complex, and might associate with the severity or outcome of severe alcoholic hepatitis (SAH). METHODS: We performed a cross-sectional study of 265 patients with SAH admitted in the Department of Hepatology, Institute of Liver and Biliary Sciences in New Delhi, India from January 2014 through January 2016. Blood samples were collected and patients were followed for 12 months or death. The molar ratios of bilirubin: albumin and albumin-bilirubin complexes were determined for a discovery cohort (30 patients who survived the study period and 60 patients who did not survive) and compared with those of 60 patients with alcoholic cirrhosis and 30 healthy individuals (controls). Optical activities of albumin-bilirubin complexes in blood samples were determined by CD spectroscopy and compared among groups. Findings were validated in a separate cohort of 150 patients with SAH from the same institute. We studied the correlation between ME and albumin binding capacity (ABiC). RESULTS: The molar ratio of bilirubin: albumin was higher in patients with SAH than with alcoholic cirrhosis or controls (P < .05). Patients with SAH had different CD spectra and higher ME than the other groups (P < .01); ME correlated with model for end-stage liver disease score (with and without Na) and discriminant function (r2 > .3; P < .01). ME values above a cut off of 1.84 mdeg predicted 3-month mortality in patients with SAH with an area under receiver operating characteristic curve of 0.87 (95% CI, 0.79-0.95), a 77% positive predictive value, and a 90% negative predictive value. The hazard ratio and concordance index of ME values for 3-month mortality in patients with SAH was 10% higher than the hazard ratio and concordance index of model for end-stage liver disease score. In patients with SAH, there was an inverse correlation between ME and ABiC (r2 > 0.7; P < .01). We observed a significant reduction in ABiC with increasing levels of bilirubin in vitro prepared albumin-bilirubin complex. CONCLUSION: In a cross-sectional study of patients with SAH, we associated ME of the albumin-bilirubin complex, measured by CD spectroscopy, with outcomes of patients with SAH. Increased loading of bilirubin on albumin could explain reduced albumin function. Bilirubin removal by albumin dialysis might benefit patients with SAH.

Mucosal Expression of Type 2 and Type 17 Immune Response Genes Distinguishes Ulcerative Colitis From Colon-Only Crohn's Disease in Treatment-Naive Pediatric Patients.

BACKGROUND & AIMS: There is controversy regarding the role of the type 2 immune response in the pathogenesis of ulcerative colitis (UC)-few data are available from treatment-naive patients. We investigated whether genes associated with a type 2 immune response in the intestinal mucosa are up-regulated in treatment-naive pediatric patients with UC compared with patients with Crohn's disease (CD)-associated colitis or without inflammatory bowel disease (IBD), and whether expression levels are associated with clinical outcomes. METHODS: We used a real-time reverse-transcription quantitative polymerase chain reaction array to analyze messenger RNA (mRNA) expression patterns in rectal mucosal samples from 138 treatment-naive pediatric patients with IBD and macroscopic rectal disease, as well as those from 49 children without IBD (controls), enrolled in a multicenter prospective observational study from 2008 to 2012. Results were validated in real-time reverse-transcription quantitative polymerase chain reaction analyses of rectal RNA from an independent cohort of 34 pediatric patients with IBD and macroscopic rectal disease and 17 controls from Cincinnati Children's Hospital Medical Center. RESULTS: We measured significant increases in mRNAs associated with a type 2 immune response (interleukin [IL]5 gene, IL13, and IL13RA2) and a type 17 immune response (IL17A and IL23) in mucosal samples from patients with UC compared with patients with colon-only CD. In a regression model, increased expression of IL5 and IL17A mRNAs distinguished patients with UC from patients with colon-only CD (P = .001; area under the receiver operating characteristic curve, 0.72). We identified a gene expression pattern in rectal tissues of patients with UC, characterized by detection of IL13 mRNA, that predicted clinical response to therapy after 6 months (odds ratio [OR], 6.469; 95% confidence interval [CI], 1.553-26.94), clinical response after 12 months (OR, 6.125; 95% CI, 1.330-28.22), and remission after 12 months (OR, 5.333; 95% CI, 1.132-25.12). CONCLUSIONS: In an analysis of rectal tissues from treatment-naive pediatric patients with IBD, we observed activation of a type 2 immune response during the early course of UC. We were able to distinguish patients with UC from those with colon-only CD based on increased mucosal expression of genes that mediate type 2 and type 17 immune responses. Increased expression at diagnosis of genes that mediate a type 2 immune response is associated with response to therapy and remission in pediatric patients with UC.

Factors That Predict High Health Care Utilization and Costs for Patients With Inflammatory Bowel Diseases.

BACKGROUND & AIMS: A subset of patients with inflammatory bowel diseases (IBD) have continuously active inflammation, leading to a high number of complications and high direct health care costs (diagnostic tests, medications, and surgeries) and indirect costs (reduced employment and productivity and fewer opportunities for activities). Identifying these high-risk patients and providing effective interventions could produce better outcomes and reduce costs. We used prior year data to create IBD risk models to predict IBD-related hospitalizations, emergency department visits, and high treatment charges (>$30,000/year) in the subsequent year. METHODS: We performed a retrospective study of medical records from all patients with IBD treated at the University of Michigan Hospital from fiscal years 2013-2015. We selected clinical variables from the prior year and tested their abilities to predict 3 adverse outcomes (IBD-related hospitalizations, emergency department visits, and treatment charges >$30,000/year) in the subsequent year. Individual patients were only included once in the data set. We created a multivariate model that was based on a 70% randomly selected cohort (1005 patients) and validated the model on the other 30% (425 patients). Logistic regression was used for bivariate and multivariate analyses. RESULTS: Factors that predicted high-cost outcomes included the presence of psychiatric illness, use of corticosteroids, use of narcotics, low levels of hemoglobin, and high numbers of IBD-related hospitalizations. In the validation cohort, the model predicted IBD-related hospitalizations, emergency department visits, and high charges in the following year with receiver operating characteristic curve values of 0.751, 0.738, and 0.744, respectively. CONCLUSIONS: We identified 5 factors that can effectively identify patients with IBD at high risk for hospitalization, emergency department visits, and high treatment charges in the next year. These patients should be closely monitored and aggressively managed.

Validation of the Simplified Criteria for the Diagnosis of Autoimmune Hepatitis in Chilean-Hispanic Patients.

INTRODUCTION AND AIM: In 2008 the International autoimmune hepatitis (AIH) Group proposed the simplified diagnostic criteria for this disease. The original cohort study was performed in 11 international centers, but validation studies are scarce in Latin-America. The aim of this study is validate these criteria in Hispanic patients. MATERIAL AND METHODS: A retrospective cohort of patients undergoing percutaneous liver biopsy and follow-up of at least 12 months was recruited from a Chilean University hospital. Patients with previous immunosuppressive therapy and liver transplant recipients were excluded. The diagnostic accuracy was analyzed using as gold standard the clinical course during long-term follow-up. Sensitivity, specificity, positive and negative predictive values (PPV and NPV) and area under the ROC curve (AUROC) were calculated. RESULTS: Four hundred eighty one patients were evaluated, 294 were included. 218 (74.15%) were female, mean age 48.5 (± 12.3) years, mean follow-up 34 (± 18) months. 66 patients had AIH or overlap syndrome (22.45%), 96 (32.65%) non-alcoholic steatohepatitis, 40 (13.61%) primary biliary cholangitis, 31 (10.54%) hepatitis C, 8 (2.72%) hepatitis B, 53 (18.02%) other etiologies. The AUROC for AIH simplified criteria was 0.976. Using a cutoff ≥ 6 and ≥ 7 points, the sensitivity was 86.4% and 54.6%; specificity, 98.7% and 99.6%; PPV, 95% and 97.3%; and NPV, 96.2% and 88.6%, respectively. CONCLUSION: Simplified criteria for the diagnosis of AIH have a high accuracy in our Chilean-Hispanic cohort. The female gender is strongly associated to AIH and could help in difficult cases. Further studies with a prospective design are necessary to confirm these observations.

Thirty-day readmission and reoperation after surgery for spinal tumors: a National Surgical Quality Improvement Program analysis.

OBJECTIVE The goal of this study was to use a large national registry to evaluate the 30-day cumulative incidence and predictors of adverse events, readmissions, and reoperations after surgery for primary and secondary spinal tumors. METHODS Data from adult patients who underwent surgery for spinal tumors (2011-2014) were extracted from the prospective National Surgical Quality Improvement Program (NSQIP) registry. Multivariable logistic regression was used to evaluate predictors of reoperation, readmission, and major complications (death, neurological, cardiopulmonary, venous thromboembolism [VTE], surgical site infection [SSI], and sepsis). Variables screened included patient age, sex, tumor location, American Society of Anesthesiologists (ASA) physical classification, preoperative functional status, comorbidities, preoperative laboratory values, case urgency, and operative time. Additional variables that were evaluated when analyzing readmission included complications during the surgical hospitalization, hospital length of stay (LOS), and discharge disposition. RESULTS Among the 2207 patients evaluated, 51.4% had extradural tumors, 36.4% had intradural extramedullary tumors, and 12.3% had intramedullary tumors. By spinal level, 20.7% were cervical lesions, 47.4% were thoracic lesions, 29.1% were lumbar lesions, and 2.8% were sacral lesions. Readmission occurred in 10.2% of patients at a median of 18 days (interquartile range [IQR] 12-23 days); the most common reasons for readmission were SSIs (23.7%), systemic infections (17.8%), VTE (12.7%), and CNS complications (11.9%). Predictors of readmission were comorbidities (dyspnea, hypertension, and anemia), disseminated cancer, preoperative steroid use, and an extended hospitalization. Reoperation occurred in 5.3% of patients at a median of 13 days (IQR 8-20 days) postoperatively and was associated with preoperative steroid use and ASA Class 4-5 designation. Major complications occurred in 14.4% of patients: the most common complications and their median time to occurrence were VTE (4.5%) at 9 days (IQR 4-19 days) postoperatively, SSIs (3.6%) at 18 days (IQR 14-25 days), and sepsis (2.9%) at 13 days (IQR 7-21 days). Predictors of major complications included dependent functional status, emergency case status, male sex, comorbidities (dyspnea, bleeding disorders, preoperative systemic inflammatory response syndrome, preoperative leukocytosis), and ASA Class 3-5 designation (p < 0.05). The median hospital LOS was 5 days (IQR 3-9 days), the 30-day mortality rate was 3.3%, and the median time to death was 20 days (IQR 12.5-26 days). CONCLUSIONS In this NSQIP analysis, 10.2% of patients undergoing surgery for spinal tumors were readmitted within 30 days, 5.3% underwent a reoperation, and 14.4% experienced a major complication. The most common complications were SSIs, systemic infections, and VTE, which often occurred late (after discharge from the surgical hospitalization). Patients were primarily readmitted for new complications that developed following discharge rather than exacerbation of complications from the surgical hospital stay. The strongest predictors of adverse events were comorbidities, preoperative steroid use, and higher ASA classification. These models can be used by surgeons to risk-stratify patients preoperatively and identify those who may benefit from increased surveillance following hospital discharge.

Validation of the hepatoma arterial embolization prognostic score in European and Asian populations and proposed modification.

BACKGROUND & AIMS: Transarterial chemoembolization (TACE) is used to treat hepatocellular carcinoma (HCC), but it is a challenge to predict patient survival. The hepatic arterial embolization prognostic (HAP) score has been shown to predict which patients will have shorter survival times and should not undergo TACE. We aimed to validate this scoring system in a prospective study of patients in Europe and Asia. METHODS: We evaluated the prognostic accuracy of the HAP score in estimating overall survival (OS) of 126 patients with HCC who received TACE in the United Kingdom or Italy (training set) from 2001 through 2013. We also analyzed data from 723 patients treated in Korea and Japan (validation set), including 79 with newly diagnosed HCC, who underwent TACE in Korea or Japan from 2004 through 2013. Response to TACE was determined based on computed tomography analysis. OS was calculated from the time of the first TACE until death or the last follow-up evaluation. RESULTS: OS was associated with hypoalbuminemia, α-fetoprotein level greater than 400 ng/mL, and tumor size greater than 7 cm at diagnosis (P < .01), but not a bilirubin level greater than 17 umol/L (P > .05), in both data sets. The lack of association between OS and bilirubin level was confirmed using receiver operating characteristic analysis. We developed a modified version of the HAP score, based on the level of albumin and α-fetoprotein and tumor size, which predicted OS with increased accuracy in the training and validation cohorts. CONCLUSIONS: In a multicenter validation study, we developed a modified version of the HAP that predicts survival of patients with HCC treated with TACE in Europe and Asia. This system might be used to identify patients with HCC most likely to benefit from TACE in clinical practice.

A MELD-based model to determine risk of mortality among patients with acute variceal bleeding.

BACKGROUND & AIMS: Patients with cirrhosis with acute variceal bleeding (AVB) have high mortality rates (15%-20%). Previously described models are seldom used to determine prognoses of these patients, partially because they have not been validated externally and because they include subjective variables, such as bleeding during endoscopy and Child-Pugh score, which are evaluated inconsistently. We aimed to improve determination of risk for patients with AVB. METHODS: We analyzed data collected from 178 patients with cirrhosis (Child-Pugh scores of A, B, and C: 15%, 57%, and 28%, respectively) and esophageal AVB who received standard therapy from 2007 through 2010. We tested the performance (discrimination and calibration) of previously described models, including the model for end-stage liver disease (MELD), and developed a new MELD calibration to predict the mortality of patients within 6 weeks of presentation with AVB. MELD-based predictions were validated in cohorts of patients from Canada (n = 240) and Spain (n = 221). RESULTS: Among study subjects, the 6-week mortality rate was 16%. MELD was the best model in terms of discrimination; it was recalibrated to predict the 6-week mortality rate with logistic regression (logit, -5.312 + 0.207 • MELD; bootstrapped R(2), 0.3295). MELD values of 19 or greater predicted 20% or greater mortality, whereas MELD scores less than 11 predicted less than 5% mortality. The model performed well for patients from Canada at all risk levels. In the Spanish validation set, in which all patients were treated with banding ligation, MELD predictions were accurate up to the 20% risk threshold. CONCLUSIONS: We developed a MELD-based model that accurately predicts mortality among patients with AVB, based on objective variables available at admission. This model could be useful to evaluate the efficacy of new therapies and stratify patients in randomized trials.

Liver stiffness-based optimization of hepatocellular carcinoma risk score in patients with chronic hepatitis B.

BACKGROUND & AIMS: CU-HCC score is accurate to predict hepatocellular carcinoma (HCC) in chronic hepatitis B (CHB) patients. However, diagnosis of cirrhosis may be incorrect based on ultrasonography, leading to some errors in HCC prediction. This study aimed to evaluate the accuracy of LSM-HCC score, refined from CU-HCC score with liver stiffness measurement (LSM) using transient elastography to predict HCC. METHODS: A prospective cohort study of 1555 consecutive CHB patients referred for transient elastography examination; 1035 and 520 patients randomly assigned to training and validation cohorts, respectively. Clinical cirrhosis of CU-HCC score was substituted by LSM and analyzed with multivariable Cox regression analysis with other parameters. RESULTS: During a mean follow-up of 69 months, 38 patients (3.7%) in the training cohort and 17 patients (3.4%) in the validation cohort developed HCC. A new LSM-HCC score composed of LSM, age, serum albumin and hepatitis B virus (HBV) DNA levels were derived, which ranges from 0 to 30. Areas under receiver operating characteristic curves of LSM-HCC score were higher than those of CU-HCC score (0.83-0.89 vs. 0.75-0.81). By applying the cutoff value of 11, the score excluded future HCC with high negative predictive value (99.4%-100%) at 5 years. CONCLUSIONS: LSM-HCC score constructed from LSM, age, serum albumin and HBV DNA level is accurate to predict HCC in CHB patients.

Invasive aspergillosis in patients with severe alcoholic hepatitis.

BACKGROUND & AIMS: Severe alcoholic hepatitis (AH) has a poor short-term prognosis. Although infections are frequent complications of AH, the incidence of invasive aspergillosis (IA) and its impact on outcome remain unknown. METHODS: We prospectively followed 94 biopsy-proven severe AH episodes for 3 months. We retrospectively reviewed our diagnosis of IA based on EORTC/MSG and AspICU criteria, except for host factors. RESULTS: Fifteen IA (6 proven, 8 probable, and 1 possible) were diagnosed after a median delay of 26 days following diagnosis of AH. The sites of infection were the lungs (n=11) and central nervous system (n=2), while IA was disseminated in 2 cases. Baseline MELD score ≥24 and ICU admission were independent risk factors for IA. Thirteen IA occurred in the context of corticosteroids, and 2 had received no specific treatment for AH. Non-response to corticosteroids at day 7 was not a risk factor for IA, but IA was associated with absence of liver improvement at day 28. Despite antifungal treatment, 3-month transplant-free survival of patients with IA was 0% compared to 53% in those without IA. IA, Lille score ≥0.45, and overt encephalopathy were independent predictors of transplant-free mortality. CONCLUSIONS: IA is a frequent complication of severe AH and carries a very high risk of mortality. Systematic screening for IA should be recommended in these patients. Further studies are needed to identify high-risk populations requiring antifungal prophylactic treatment.

Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease.

BACKGROUND & AIMS: With the increasing prevalence of liver disease worldwide, there is an urgent clinical need for reliable methods to diagnose and stage liver pathology. Liver biopsy, the current gold standard, is invasive and limited by sampling and observer dependent variability. In this study, we aimed to assess the diagnostic accuracy of a novel magnetic resonance protocol for liver tissue characterisation. METHODS: We conducted a prospective study comparing our magnetic resonance technique against liver biopsy. The individual components of the scanning protocol were T1 mapping, proton spectroscopy and T2* mapping, which quantified liver fibrosis, steatosis and haemosiderosis, respectively. Unselected adult patients referred for liver biopsy as part of their routine care were recruited. Scans performed prior to liver biopsy were analysed by physicians blinded to the histology results. The associations between magnetic resonance and histology variables were assessed. Receiver-operating characteristic analyses were also carried out. RESULTS: Paired magnetic resonance and biopsy data were obtained in 79 patients. Magnetic resonance measures correlated strongly with histology (r(s)=0.68 p<0.0001 for fibrosis; r(s)=0.89 p<0.001 for steatosis; r(s)=-0.69 p<0.0001 for haemosiderosis). The area under the receiver operating characteristic curve was 0.94, 0.93, and 0.94 for the diagnosis of any degree of fibrosis, steatosis and haemosiderosis respectively. CONCLUSION: The novel scanning method described here provides high diagnostic accuracy for the assessment of liver fibrosis, steatosis and haemosiderosis and could potentially replace liver biopsy for many indications. This is the first demonstration of a non-invasive test to differentiate early stages of fibrosis from normal liver.

Outcome of patients with cirrhosis requiring mechanical ventilation in ICU.

INTRODUCTION: Mortality rate of patients with cirrhosis admitted to the intensive care unit (ICU) and requiring mechanical ventilation varies between 60 and 91%. The aim of our study is to assess the prognosis of these patients, their 1-year outcome and to analyze predictive factors of long-term mortality. METHODS: From May 2005 to May 2011, we studied 246 consecutive patients with cirrhosis requiring mechanical ventilation either at admission or during their ICU stay. RESULTS: Alcohol was the most common etiology of the cirrhosis (69%). Bleeding related to portal hypertension (30%) and severe sepsis (33%) were the most common reasons for admission. ICU and hospital mortality were respectively 65.9% and 70.3%. Prognostic severity scores, the need for other organ support therapy, infection, and total bilirubin value at ICU admission were significantly associated with ICU mortality. Eighty-four patients (34.1%) were discharged from the ICU. Among these patients, the one-year survival was only of 32%. Logistic regression analysis, using survival at one year as the endpoint, identified two independent risk factors: the length of ventilation (odds ratio [OR] = 1.1; 95% CI, 1.0-1.2; p = 0.02) and total bilirubin at ICU discharge (OR = 1.3; 95% CI, 1.1-1.5; p = 0.006). CONCLUSION: Patients with cirrhosis admitted to the liver ICU and who required mechanical ventilation have a poor prognosis with a 1-year mortality of 89%. At ICU discharge, a total bilirubin level higher than 64.5 µmol/L and length of ventilation higher than 9 days could help the hepatologists to identify patients at risk of death in the year following the ICU discharge.