Evaluation of D-dimer and prothrombin time in alcohol related liver cirrhosis with comparison of machine learning analyses.

OBJECTIVES: Liver cirrhosis (LC) can be caused by obesity, alcohol consumption, viral infection, and autoimmune disease. Early diagnosis and management of LC is important for patient quality of life. Non-invasive diagnostic methods are useful for predicting the current status and mortality risk of LC. The purpose of this study is to identify relevant diagnostic factors measured in routine laboratory test of alcohol-related liver cirrhosis (ALC) patients. METHODS: This study analyzed data from 127 patients with ALC, including their laboratory test results and clinical information, including coagulation parameters, hematologic parameters, and biochemical parameters. These data were used to compare the performance of the prediction models from three machine learning algorithms including K-nearest neighbor (KNN), support vector machine (SVM), and random forest (RF). RESULTS: Higher Model for End-stage Liver Disease (MELD) score were associated with prothrombin time (PT) and D-dimer. Logistic and multiple linear regression analyses revealed significant factors predicting mortality in the MELD group. Machine learning approaches were used to predict death in ALC patients using some laboratory parameters associated with mortality. The prediction model based on SVM exhibited better prediction performance than others. CONCLUSION: PT and D-dimer were the factors that were most strongly associated with 90-day mortality, and machine learning methods can create prediction models with good predictive power.

Classification Model for Diabetic Foot, Necrotizing Fasciitis, and Osteomyelitis.

Diabetic foot ulcers (DFUs) and their life-threatening complications, such as necrotizing fasciitis (NF) and osteomyelitis (OM), increase the healthcare cost, morbidity and mortality in patients with diabetes mellitus. While the early recognition of these complications could improve the clinical outcome of diabetic patients, it is not straightforward to achieve in the usual clinical settings. In this study, we proposed a classification model for diabetic foot, NF and OM. To select features for the classification model, multidisciplinary teams were organized and data were collected based on a literature search and automatic platform. A dataset of 1581 patients (728 diabetic foot, 76 NF, and 777 OM) was divided into training and validation datasets at a ratio of 7:3 to be analyzed. The final prediction models based on training dataset exhibited areas under the receiver operating curve (AUC) of the 0.80 and 0.73 for NF model and OM model, respectively, in validation sets. In conclusion, our classification models for NF and OM showed remarkable discriminatory power and easy applicability in patients with DFU.

Identification of the robust predictor for sepsis based on clustering analysis.

Sepsis is a life-threatening disorder with high incidence and mortality rate. However, the early detection of sepsis is challenging due to lack of specific marker and various etiology. This study aimed to identify robust risk factors for sepsis via cluster analysis. The integrative task of the automatic platform (i.e., electronic medical record) and the expert domain was performed to compile clinical and medical information for 2,490 sepsis patients and 16,916 health check-up participants. The subjects were categorized into 3 and 4 groups based on seven clinical and laboratory markers (Age, WBC, NLR, Hb, PLT, DNI, and MPXI) by K-means clustering. Logistic regression model was performed for all subjects including healthy control and sepsis patients, and cluster-specific cases, separately, to identify sepsis-related features. White blood cell (WBC), well-known parameter for sepsis, exhibited the insignificant association with the sepsis status in old age clusters (K3C3 and K4C3). Besides, NLR and DNI were the robust predictors in all subjects as well as three or four cluster-specific subjects including K3C3 or K4C3. We implemented the cluster-analysis for real-world hospital data to identify the robust predictors for sepsis, which could contribute to screen likely overlooked and potential sepsis patients (e.g., sepsis patients without WBC count elevation).

A New Bacterial Growth Graph Pattern Analysis to Improve Positive Predictive Value of Continuous Monitoring Blood Culture System.

False positive signals (FPSs) of continuous monitoring blood culture system (CMBCS) cause delayed reporting time and increased laboratory cost. This study aimed to analyze growth graphs digitally in order to identify specific patterns of FPSs and true positive signals (TPSs) and to find the method for improving positive predictive value (PPV) of FPS and TPS. 606 positive signal samples from the BACTEC FX (BD, USA) CMBCS with more than one hour of monitoring data after positive signal were selected, and were classified into FPS and TPS groups using the subculture results. The pattern of bacterial growth graph was analyzed in two steps: the signal stage recorded using the monitoring data until positive signal and the post-signal stage recorded using one additional hour of monitoring data gained after the positive signal. The growth graph before the positive signal consists of three periods; initial decline period, stable period, and steeping period. Signal stage analyzed initial decline period and stable period, and classified the graphs as standard, increasing, decreasing, irregular, or defective pattern, respectively. Then, all patterns were re-assigned as confirmed or suspicious pattern in the post-signal stage. Standard, increasing, and decreasing patterns with both initial decline period and stable period are typical patterns; irregular patterns lacking a smooth stable period and defective patterns without an initial decline period are false positive patterns. The false positive patterns have 77.2% of PPV for FPS. The confirmed patterns, showing a gradually increasing fluorescence level even after positive signal, have 97.0% of PPV for TPS.

The characteristics of transferrin variants by carbohydrate-deficient transferrin tests using capillary zone electrophoresis.

BACKGROUND: Transferrin is the major plasma transport protein for iron. We aimed to investigate the characteristics of transferrin variant by carbohydrate-deficient transferrin (CDT) test using capillary zone electrophoresis. METHODS: We retrospectively analyzed the CDT tests of 2449 patients from March 2009 to May 2017 at a tertiary hospital in Korea. CDT was quantified using a Capillarys 2 system (Sebia, Lisses, France) by capillary zone electrophoresis. The characteristics of variant transferrin patterns using electropherogram of CDT tests were analyzed. RESULTS: Seventy-seven (3.1%) patients were classified as variant transferrin. Mean age of these patients was 51.8 years, and the male-to-female ratio was 3.5:1. The most common variants were the BC variants (n = 37), followed by the CD variants (n = 27), unclear patterns (n = 7), BD variants (n = 3), CC variants (n = 2), misclassification (n = 1). In the variant Tf group, the most common disease was alcoholic liver cirrhosis (n = 22, 28.6%), followed by the toxic effects of substances (n = 17, 22.1%), and mental and behavioral disorders attributable to alcohol (n = 11, 14.3%). Nonvariant group showed a predominance of the toxic substance effects (n = 880, 37.1%), a personal history of suicide attempts (n = 634, 26.7%), and mental and behavioral disorders due to alcohol (n = 336, 14.2%). CONCLUSION: We analyzed the basic characteristics of variant transferrin by CDT tests using capillary zone electrophoresis. The prevalence of variant transferrin was 3.1% of the study subjects. Male patients, alcohol abusers, and liver cirrhosis patients predominated in the variant transferrin population. Further prospective studies are warranted to elucidate variant transferrin in clinical practice.

Development of a Rapid Reverse Blot Hybridization Assay for Detection of Clinically Relevant Antibiotic Resistance Genes in Blood Cultures Testing Positive for Gram-Negative Bacteria.

Rapid and accurate identification of the causative pathogens of bloodstream infections is crucial for the prompt initiation of appropriate antimicrobial therapy to decrease the related morbidity and mortality rates. The aim of this study was to evaluate the performance of a newly developed PCR-reverse blot hybridization assay (REBA) for the rapid detection of Gram-negative bacteria (GNB) and their extended-spectrum ß-lactamase (ESBL), AmpC ß-lactamase, and carbapenemase resistance genes directly from the blood culture bottles. The REBA-EAC (ESBL, AmpC ß-lactamase, carbapenemase) assay was performed on 327 isolates that were confirmed to have an ESBL producer phenotype, 200 positive blood culture (PBCs) specimens, and 200 negative blood culture specimens. The concordance rate between the results of REBA-EAC assay and ESBL phenotypic test was 94.2%. The sensitivity, specificity, positive predictive value, and negative predictive value of the REBA-EAC assay for GNB identification in blood culture specimens were 100% (95% CI 0.938-1.000, P < 0.001), 100% (95% CI 0.986-1.000, P < 0.001), 100% (95% CI 0.938-1.000, P < 0.001), and 100% (95% CI 0.986-1.000, P < 0.001), respectively. All 17 EAC-producing GNB isolates from the 73 PBCs were detected by the REBA-EAC assay. The REBA-EAC assay allowed easy differentiation between EAC and non-EAC genes in all isolates. Moreover, the REBA-EAC assay was a rapid and reliable method for identifying GNB and their ß-lactamase resistance genes in PBCs. Thus, this assay may provide essential information for accelerating therapeutic decisions to achieve earlier appropriate antibiotic treatment during the acute phase of bloodstream infection.

Scoring system for detecting spurious hemolysis in anticoagulated blood specimens.

BACKGROUND: The identification of in vitro hemolysis (IVH) using a hematology analyzer is challenging because centrifugation of the specimens cannot be performed for cell counts. In the present study, we aimed to develop a scoring system to help identify the presence of hemolysis in anticoagulated blood specimens. METHODS: Thirty-seven potassium EDTA anticoagulated blood specimens were obtained, and each specimen was divided into 3 aliquots (A, B, and C). Aliquots B and C were mechanically hemolyzed by aspirating 2 and 5 times, respectively, using a 27-gauge needle and then tested; aliquot A was analyzed immediately without any hemolysis. After the cells were counted, aliquots B and C were centrifuged and the supernatants were tested for the hemolytic index and lactate dehydrogenase levels. RESULTS: The 4 hematologic parameters were selected and scored from 0 to 3 as follows:< 34.0, 34.0-36.2, 36.3-38.4, and ≥38.5 for mean cell hemoglobin concentration (MCHC, g/dL); <0.02, 0.02, 0.03, and ≥0.04 for red blood cell ghosts (10(12)/L); <0.13, 0.13-0.38, 0.39-1.30, and ≥1.31 for difference value (g/dL) of measured hemoglobin and calculated hemoglobin; and <0.26, 0.26-0.95, 0.96-3.34, and ≥3.35 for difference value (g/dL) of MCHC and cell hemoglobin concentration mean. The hemolysis score was calculated by adding all the scores from the 4 parameters. At the cutoff hemolysis score of 3, the IVH of aliquots B and C were detected as 64.9% and 91.9%, respectively. CONCLUSIONS: The scoring system might provide effective screening for detecting spurious IVH.