Learning From Noisy Labels With Deep Neural Networks: A Survey.

Deep learning has achieved remarkable success in numerous domains with help from large amounts of big data. However, the quality of data labels is a concern because of the lack of high-quality labels in many real-world scenarios. As noisy labels severely degrade the generalization performance of deep neural networks, learning from noisy labels (robust training) is becoming an important task in modern deep learning applications. In this survey, we first describe the problem of learning with label noise from a supervised learning perspective. Next, we provide a comprehensive review of 62 state-of-the-art robust training methods, all of which are categorized into five groups according to their methodological difference, followed by a systematic comparison of six properties used to evaluate their superiority. Subsequently, we perform an in-depth analysis of noise rate estimation and summarize the typically used evaluation methodology, including public noisy datasets and evaluation metrics. Finally, we present several promising research directions that can serve as a guideline for future studies.

Comparison of mortality among hemorrhage-control methods performed for hemodynamically unstable patients with traumatic pelvic fractures: A multi-center study.

BACKGROUND: /Objective: We aimed to analyze the effects of hemorrhage control methods on the mortality of patients with hemodynamic instability due to pelvic fracture and investigate independent mortality risk factors in these patients. METHODS: Ninety-seven pelvic bone fracture patients with hemodynamic instability who visited the emergency departments of two university hospitals over 5 years were enrolled. These patients were categorized based on 28-day mortality (survival group) and acute hemorrhage mortality (non-survival group). Forty-seven patients (48.5%) underwent pelvic angiography; 45 (46.4%), pre-peritoneal pelvic packing; and 19 (19.6%), external fixation. RESULTS: Differences in hemorrhage control methods did not significantly affect mortality. However, there was a significant difference in mortality between the groups with and without hemorrhage control methods. Multivariate logistic regression analysis revealed that patient age, trauma and injury severity score (probability of survival), and blood transfusion amount within 24 h were independent risk factors for 28-day mortality. Meanwhile, patient age, Glasgow coma scale (GCS) score, systolic blood pressure (SBP), and blood transfusion amount within 24 h were independent risk factors for mortality due to acute hemorrhage. CONCLUSION: Rapid and appropriate application of hemorrhage control methods can reduce acute hemorrhage-related mortality in hemodynamically unstable patients with pelvic fractures. Moreover, none of the hemorrhage control methods were superior for the decreasing mortality rate in these patients.

Use of venovenous extracorporeal membrane oxygenation in trauma patients with severe adult respiratory distress syndrome: A retrospective study.

BACKGROUND: The use of extracorporeal membrane oxygenation (ECMO) has increased, although its survival benefit in trauma patients with severe adult respiratory distress syndrome (ARDS) remains controversial. We investigated the effect of veno-venous (VV)-ECMO on the clinical outcomes of trauma patients with severe ARDS. METHODS: This was a retrospective study at a single center comprising trauma patients admitted between January 2013 and December 2017, diagnosed with severe ARDS using the Berlin definition (PaO2/FiO2 ratio ⩽100), in the 7 days following trauma. Patients were managed with VV-ECMO or conventional mechanical ventilation (CMV). The primary outcome was in-hospital mortality (mortality at 60 days); secondary outcomes comprised 28-day mortality, hospital length of stay (LOS), intensive care unit (ICU) LOS, ICU-free days, duration of mechanical ventilation (MV), and MV-free days. Propensity score matching was performed to adjust for the baseline differences. RESULTS: Sixteen patients (22.5%) were managed with VV-ECMO and 55 were managed with CMV. After matching, the in-hospital mortality rate (43.8% vs 53.1%; p = 0.760), 28-day mortality rate (37.5% vs 31.3%; p = 0.750), median hospital LOS (39.5 vs 36.5 days; p = 0.533), ICU-free days (0 vs 0 days; p = 0.241), and MV-free days (0 vs 0 days; p = 0.272) did not significantly differ between the VV-ECMO and CMV groups. CONCLUSION: In-hospital mortality (mortality at 60 days) did not differ significantly between the VV-ECMO and CMV groups. Although the safety of ECMO in trauma patients requires further investigation, VV-ECMO may be considered as a rescue therapy.

Age-related changes in mean corpuscular volumes in patients without anaemia: An analysis of large-volume data from a single institute.

Although the mean corpuscular volume (MCV) has been associated with various diseases, these associations in relation to the age-related trends in MCV remain unclear. Therefore, we used a dataset with over one million values to identify the relationship between ageing and MCV changes. All laboratory data obtained between November 1998 and November 2019 at Chungbuk National University Hospital were retrospectively collected. After excluding cases with missing values for individual complete blood count parameters, outlier MCV values, and ages less than 1 year and more than 88 years, 977,335 MCV values were obtained from 309,393 patients. Principal component analysis of blood components with ages and analysis of the median value changes for each blood component across decade-wise age groups were conducted to identify relationships between ageing and changes in blood components. The median values of MCV showed gradual increments with age. The linear relationship for patients aged 1-25 years had a larger slope than that for patients aged 26-88 years. For MCV, the equation for patients aged 1-25 years was 0.40*(age) + 81.24 in females and 0.45*(age) + 79.58 in males. The equation for patients aged 26-90 years was 0.04*(age) + 88.97 in females and 0.06*age + 88.30 in males. Among patients aged >40 years, the MCV value was higher in men than in women. Analysis of a large dataset showed that the MCV gradually increased with age and the linear relationship differed between patients aged 1-25 and 26-88 years.

Can Robots Help Working Parents with Childcare? Optimizing Childcare Functions for Different Parenting Characteristics.

Is it true that parents always prioritize educational effectiveness when selecting childcare services? The current study identified the potential requirements of dual-income parents toward social robots' diverse childcare functions (e.g., socialization, education, entertainment, and consultation). The results revealed that parental attitudes toward robots were made more positive by all the childcare functions of robots except for their educational features. Furthermore, parents' expectations of childcare functions varied based on their parenting characteristics. Spectral clustering analysis identified distinctive parenting styles (e.g., family-oriented, work-oriented, noninterventional, and dominant), and multigroup structural equation modeling suggested that the impact of robots' socialization function was significant in all parent groups, while other childcare functions exerted limited influence according to specific parenting styles. In addition, children's characteristics were found to alter parents' preferences for each childcare function. These results offer practical implications for the early adoption of childcare robots through predetermining parents' acceptability based on their specific parenting characteristics.

A nomogram to predict arterial bleeding in patients with pelvic fractures after blunt trauma: a retrospective cohort study.

BACKGROUND: Pelvic bone fractures are one of the biggest challenges faced by trauma surgeons. Especially, the presence of bleeding and hemodynamic instability features is associated with high morbidity and mortality in patients with pelvic fractures. However, prediction of the occurrence of arterial bleeding causing massive hemorrhage in patients with pelvic fractures is difficult. Therefore, the aim of this study was to develop a nomogram to predict arterial bleeding in patients with pelvic bone fractures after blunt trauma. METHODS: The medical records of 1404 trauma patients treated between January 2013 and August 2017 were retrospectively reviewed. Patients older than 15 years with a pelvic fracture due to blunt trauma were enrolled (n = 148). The pelvic fracture pattern on anteroposterior radiography was classified according to the Orthopedic Trauma Association/Arbeitsgemeinschaft fur Osteosynthesefragen (OTA/AO) system. Multivariable logistic regression modeling was used to determine the independent risk factors for arterial bleeding. A nomogram was constructed based on the identified risk factors. RESULTS: The most common pelvic fracture pattern was type A (58.8%), followed by types B (34.5%) and C (6.7%). Of the 148 patients, 28 (18.9%) showed pelvic arterial bleeding on contrast-enhanced computed tomography or angiography, or in the operative findings. The independent risk factors for arterial bleeding were a type B or C pelvic fracture pattern, body temperature < 36 °C, and serum lactate level > 3.4 mmol/L. A nomogram was developed using these three parameters, along with a systolic blood pressure < 90 mmHg. The area under the receiver operating characteristic curve of the predictive model for discrimination was 0.8579. The maximal Youden index was 0.1527, corresponding to a cutoff value of 68.65 points, which was considered the optimal cutoff value for predicting the occurrence of arterial bleeding in patients with pelvic bone fractures. CONCLUSIONS: The developed nomogram, which was based on the initial clinical findings identifying risk factors for arterial bleeding, is expected to be helpful in rapidly establishing a treatment plan and improving the prognosis for patients with pelvic bone fractures.

DTranNER: biomedical named entity recognition with deep learning-based label-label transition model.

BACKGROUND: Biomedical named-entity recognition (BioNER) is widely modeled with conditional random fields (CRF) by regarding it as a sequence labeling problem. The CRF-based methods yield structured outputs of labels by imposing connectivity between the labels. Recent studies for BioNER have reported state-of-the-art performance by combining deep learning-based models (e.g., bidirectional Long Short-Term Memory) and CRF. The deep learning-based models in the CRF-based methods are dedicated to estimating individual labels, whereas the relationships between connected labels are described as static numbers; thereby, it is not allowed to timely reflect the context in generating the most plausible label-label transitions for a given input sentence. Regardless, correctly segmenting entity mentions in biomedical texts is challenging because the biomedical terms are often descriptive and long compared with general terms. Therefore, limiting the label-label transitions as static numbers is a bottleneck in the performance improvement of BioNER. RESULTS: We introduce DTranNER, a novel CRF-based framework incorporating a deep learning-based label-label transition model into BioNER. DTranNER uses two separate deep learning-based networks: Unary-Network and Pairwise-Network. The former is to model the input for determining individual labels, and the latter is to explore the context of the input for describing the label-label transitions. We performed experiments on five benchmark BioNER corpora. Compared with current state-of-the-art methods, DTranNER achieves the best F1-score of 84.56% beyond 84.40% on the BioCreative II gene mention (BC2GM) corpus, the best F1-score of 91.99% beyond 91.41% on the BioCreative IV chemical and drug (BC4CHEMD) corpus, the best F1-score of 94.16% beyond 93.44% on the chemical NER, the best F1-score of 87.22% beyond 86.56% on the disease NER of the BioCreative V chemical disease relation (BC5CDR) corpus, and a near-best F1-score of 88.62% on the NCBI-Disease corpus. CONCLUSIONS: Our results indicate that the incorporation of the deep learning-based label-label transition model provides distinctive contextual clues to enhance BioNER over the static transition model. We demonstrate that the proposed framework enables the dynamic transition model to adaptively explore the contextual relations between adjacent labels in a fine-grained way. We expect that our study can be a stepping stone for further prosperity of biomedical literature mining.

Stable Subloop Behavior in Ferroelectric Si-Doped HfO2.

The recent demand for analogue devices for neuromorphic applications requires modulation of multiple nonvolatile states. Ferroelectricity with multiple polarization states enables neuromorphic applications with various architectures. However, deterministic control of ferroelectric polarization states with conventional ferroelectric materials has been met with accessibility issues. Here, we report unprecedented stable accessibility with robust stability of multiple polarization states in ferroelectric HfO2. Through the combination of conventional voltage measurements, hysteresis temperature dependence analysis, piezoelectric force microscopy, first-principles calculations, and Monte Carlo simulations, we suggest that the unprecedented stability of intermediate states in ferroelectric HfO2 is due to the small critical volume size for nucleation and the large activation energy for ferroelectric dipole flipping. This work demonstrates the potential of ferroelectric HfO2 for analogue device applications enabling neuromorphic computing.

Association of Initial Low Serum Selenium Level with Infectious Complications and 30-Day Mortality in Multiple Trauma Patients.

Low serum selenium levels are commonly observed in critically injured multiple trauma patients. This study aimed to identify the association between initial serum selenium levels and in-hospital infectious complications in multiple trauma patients. We retrospectively reviewed multiple trauma patients admitted between January 2015 and November 2017. We selected 135 patients whose serum selenium levels were checked within 48 h of admission. Selenium deficiency was defined as a serum selenium level <70 ng/mL. Survival analyses of selenium deficiency and 30-day mortality were performed. Multivariate logistic regression analysis was performed to identify the association between initial serum selenium level and in-hospital infectious complications. Thirty-day mortality (8.3% vs. 0.0%; p = 0.018) and incidence rates of pneumonia (66.7% vs. 28.3%; p < 0.001) and infectious complications (83.3% vs. 46.5%; p < 0.001) were higher in patients with selenium deficiency than in patients without selenium deficiency. Kaplan-Meier survival cures also showed similar results (log rank test, p = 0.021). Of 135 patients, 76 (56.3%) experienced at least one infectious complication during admission. High injury severity score (ISS, odds ratio (OR) 1.065, 95% confidence interval (CI) 1.024-1.108; p = 0.002) and selenium deficiency (OR 3.995, 95% CI 1.430-11.156; p = 0.008) increased the risk of in-hospital infectious complications in multiple trauma patients. Patients with selenium deficiency showed higher 30-day mortality and higher risks of pneumonia and infectious complications.

Laparoscopic appendectomy versus open appendectomy for suspected appendicitis during pregnancy: a systematic review and updated meta-analysis.

BACKGROUND: Recently, laparoscopic appendectomies (LAs) have been widely performed instead of open appendectomies (OAs) during pregnancy. However, concerns about the safety of LA during pregnancy remain. This systematic review and meta-analysis aimed to evaluate the current evidence relating to the safety of LA versus OA for suspected appendicitis during pregnancy. METHODS: Comprehensive literature searches were conducted using the PubMed, EMBASE, and Cochrane Library databases to identify articles describing LA versus OA in pregnancy, without restrictions regarding the publication date. The primary endpoints were fetal loss and preterm delivery. RESULTS: After screening 801 studies, 22 comparative cohort studies were included in the analysis, which involved 4694 women, of whom 905 underwent LAs and 3789 underwent OAs. Fetal loss was significantly higher among those who underwent LAs compared with those who underwent OAs, and the pooled odds ratio (OR) was 1.72 (95% confidence interval [CI]: 1.22-2.42) without heterogeneity. The sensitivity analysis showed that the effect size was influenced by one of the studies, because its removal resulted in there being no significant difference between LA and OA with respect to the risk of fetal loss (OR 1.163, 95% CI: 0.68-1.99; P = 0.581). A significant difference was not evident between LA and OA with respect to preterm delivery (OR 0.76, 95% CI: 0.51-1.15), a result that did not change following the sensitivity analysis. The patients who underwent LA had shorter hospital stays (mean difference - 1.01, 95% CI: -1.61--0.41) and a lower wound infection risk (OR 0.40, 95% CI: 0.21-0.76) compared with those who underwent OA. CONCLUSION: It is not reasonable to conclude that LA in pregnant women might be associated with a greater risk of fetal loss. The difference between LA and OA with respect to preterm delivery was not significant.

Antibiotic use in patients with abdominal injuries: guideline by the Korean Society of Acute Care Surgery.

PURPOSE: A task force appointed by the Korean Society of Acute Care Surgery reviewed previously published guidelines on antibiotic use in patients with abdominal injuries and adapted guidelines for Korea. METHODS: Four guidelines were assessed using the Appraisal of Guidelines for Research and Evaluation II instrument. Five topics were considered: indication for antibiotics, time until first antibiotic use, antibiotic therapy duration, appropriate antibiotics, and antibiotic use in abdominal trauma patients with hemorrhagic shock. RESULTS: Patients requiring surgery need preoperative prophylactic antibiotics. Patients who do not require surgery do not need antibiotics. Antibiotics should be administered as soon as possible after injury. In the absence of hollow viscus injury, no additional antibiotic doses are needed. If hollow viscus injury is repaired within 12 hours, antibiotics should be continued for ≤ 24 hours. If hollow viscus injury is repaired after 12 hours, antibiotics should be limited to 7 days. Antibiotics can be administered for ≥7 days if hollow viscus injury is incompletely repaired or clinical signs persist. Broad-spectrum aerobic and anaerobic coverage antibiotics are preferred as the initial antibiotics. Second-generation cephalosporins are the recommended initial antibiotics. Third-generation cephalosporins are alternative choices. For hemorrhagic shock, the antibiotic dose may be increased twofold or threefold and repeated after transfusion of every 10 units of blood until there is no further blood loss. CONCLUSION: Although this guideline was drafted through adaptation of other guidelines, it may be meaningful in that it provides a consensus on the use of antibiotics in abdominal trauma patients in Korea.

Factors predicting the need for hemorrhage control intervention in patients with blunt pelvic trauma: a retrospective study.

BACKGROUND: Blunt pelvic injuries are often associated with pelvic fractures and injuries to the rectum and genitourinary tract. Pelvic fractures can lead to life-threatening hemorrhage, which is a common cause of morbidity and mortality in trauma. Thus, early identification of patients with pelvic fractures at risk severe bleeding requiring urgent hemorrhage control is crucial. This study aimed to investigate early factors predicting the need for hemorrhage control in blunt pelvic trauma. METHODS: The medical records of 1760 trauma patients were reviewed retrospectively between January 2013 and June 2018. We enrolled 187 patients with pelvic fracture due to blunt trauma who were older than 15 years. The pelvic fracture pattern was classified according to the Orthopedic Trauma Association/Arbeitsgemeinschaft fur Osteosynthesefragen (OTA/AO) classification. A multivariate logistic regression model was used to determine independent predictors of the need for pelvic hemorrhage control intervention. RESULTS: The most common pelvic fracture pattern was type A (54.5%), followed by types B (36.9%) and C (8.6%). Of 187 patients, 48 (25.7%) required pelvic hemorrhage control intervention. Hemorrhage control interventions were most frequently performed in patients with type B fractures (54.2%). Multivariate logistic regression analysis revealed that type B (odds ratio [OR] = 4.024, 95% confidence interval [CI] = 1.666-9.720, p = 0.002) and C (OR = 7.077, 95% CI = 1.781-28.129, p = 0.005) fracture patterns, decreased body temperature (OR = 2.275, 95% CI = 0.134-0.567, p < 0.001), and elevated serum lactate level (OR = 1.234, 95% CI = 1.061-1.435, p = 0.006) were factors predicting the need for hemorrhage control intervention in patients with blunt pelvic trauma. CONCLUSION: Patients with type B and C fracture patterns on the OTA/AO classification, hypothermia, or an elevated serum lactate level are at risk for bleeding and require pelvic hemorrhage control intervention.

Association of Inadequate Caloric Supplementation with 30-Day Mortality in Critically Ill Postoperative Patients with High Modified NUTRIC Score.

Modified NUTRIC (mNUTRIC) score is a useful assessment tool to determine the risk of malnutrition in patients on mechanical ventilation (MV). We identified associations between postoperative calorie adequacy, 30-day mortality, and surgical outcomes in patients with high mNUTRIC scores. Medical records of 272 patients in the intensive care unit who required MV support for >24 h after emergency gastro-intestinal (GI) surgery between January 2007 and December 2017 were reviewed. Calorie adequacy in percentage (Calorie intake in 5 days ÷ Calorie requirement for 5 days × 100) was assessed in patients with high (5⁻9) and low (0⁻4) mNUTRIC scores. In the high mNUTRIC score group, patients with inadequate calorie supplementation (calorie adequacy <70%) had higher 30-day mortality than those with adequate supplementation (31.5% vs. 11.1%; p = 0.010); this was not observed in patients with low mNUTRIC scores. This result was also confirmed through Kaplan⁻Meier survival curve (p = 0.022). Inadequate calorie supplementation in the high mNUTRIC score group was not associated with Intra-abdominal infection (p = 1.000), pulmonary complication (p = 0.695), wound complication (p = 0.407), postoperative leakage (p = 1.000), or infections (p = 0.847). Inadequate calorie supplementation after GI surgery was associated with higher 30-day mortality in patients with high mNUTRIC scores. Therefore, adequate calorie supplementation could contribute to improved survival of critically ill postoperative patients with high risk of malnutrition.

Predictors of early weaning failure from mechanical ventilation in critically ill patients after emergency gastrointestinal surgery: A retrospective study.

Mechanical ventilation (MV) is the most common therapeutic modality used for critically ill patients. However, prolonged MV is associated with high morbidity and mortality. Therefore, it is important to avoid both premature extubation and unnecessary prolongation of MV. Although some studies have determined the predictors of early weaning success and failure, only a few have investigated these factors in critically ill surgical patients who require postoperative MV. The aim of this study was to evaluate predictors of early weaning failure from MV in critically ill patients who had undergone emergency gastrointestinal (GI) surgery.The medical records of 3327 adult patients who underwent emergency GI surgery between January 2007 and December 2016 were reviewed retrospectively. Clinical and laboratory parameters before surgery and within 2 days postsurgery were investigated.This study included 387 adult patients who required postoperative MV. A low platelet count (adjusted odds ratio [OR]: 0.995; 95% confidence interval [CI]: 0.991-1.000; P = .03), an elevated delta neutrophil index (DNI; adjusted OR: 1.025; 95% CI: 1.005-1.046; P = .016), a delayed spontaneous breathing trial (SBT; adjusted OR: 14.152; 95% CI: 6.571-30.483; P < .001), and the presence of postoperative shock (adjusted OR: 2.436; 95% CI: 1.138-5.216; P = .022) were shown to predict early weaning failure from MV in the study population.Delayed SBT, a low platelet count, an elevated DNI, and the presence of postoperative shock are independent predictors of early weaning failure from MV in critically ill patients after emergency GI surgery.

Severe persistent hypocholesterolemia after emergency gastrointestinal surgery predicts in-hospital mortality in critically ill patients with diffuse peritonitis.

BACKGROUND: Plasma cholesterol acts as a negative acute phase reactant. Total cholesterol decreases after surgery and in various pathological conditions, including trauma, sepsis, burns, and liver dysfunction. This study aimed to determine whether hypocholesterolemia after emergency gastrointestinal (GI) surgery is associated with in-hospital mortality in patients with diffuse peritonitis. METHODS: The medical records of 926 critically ill patients who had undergone emergency GI surgery for diffuse peritonitis, between January 2007 and December 2015, were retrospectively analyzed. The integrated areas under the curve (iAUCs) were calculated to compare the predictive accuracy of total cholesterol values from postoperative days (PODs) 0, 1, 3, and 7. Cox proportional hazard regression modeling was performed for all possible predictors identified in the univariate and multivariable analyses. RESULTS: The total cholesterol level measured on POD 7 had the highest iAUC (0.7292; 95% confidence interval, 0.6696-0.7891) and was significantly better at predicting in-hospital mortality than measurements on other days. The optimal total cholesterol cut-off value for predicting in-hospital mortality was 61 mg/dL and was determined on POD 7. A Cox proportional hazard regression analysis revealed that a POD 7 total cholesterol level < 61 mg/dL was an independent predictor of in-hospital mortality after emergency GI surgery (hazard ratio, 3.961; 95% confidence interval, 1.786-8.784). CONCLUSION: Severe persistent hypocholesterolemia (<61 mg/dL) on POD 7 independently predicted in-hospital mortality, after emergency GI surgery, in critically ill patients with diffuse peritonitis.

Addition of lactic acid levels improves the accuracy of quick sequential organ failure assessment in predicting mortality in surgical patients with complicated intra-abdominal infections: a retrospective study.

Background: The quick sequential organ failure assessment (qSOFA) alone has a poor sensitivity for predicting mortality in patients with complicated intra-abdominal infections, and plasma lactate levels have been shown to have a strong association with mortality in critically ill patients. Therefore, this study aimed to compare the performance of qSOFA with a score derived from a combination of qSOFA and serum lactate levels for predicting mortality in surgical patients with complicated intra-abdominal infections. Methods: This retrospective study was performed at a university hospital. The medical records of 457 patients who presented to the emergency department (ED) between January 2008 and December 2016 and required emergency gastrointestinal surgery for a complicated intra-abdominal infection were reviewed retrospectively. qSOFA criteria, sequential organ failure assessment (SOFA) scores, and plasma lactate levels during their ED stay were collected. We performed area under receiver operating characteristic (AUROC) curve and sensitivity analysis to compare the performance of qSOFA alone with that of a score derived from the use of a combination of the qSOFA and lactate levels for predicting patient mortality. Results: Fifty patients (10.9%) died during hospitalization. The combined qSOFA and lactate level score was superior to qSOFA alone (AUROC = 0.754 vs. 0.717, p = 0.039, respectively) and comparable to the full SOFA score (AUROC = 0.754 vs. 0.795, p = 0.127, respectively) in predicting mortality. Sensitivity and specificity of qSOFA alone were 46 and 86%, respectively, and those of the combined score were 72 and 73%, respectively (p < 0.001). Conclusion: A score derived from the qSOFA and serum lactate levels had better predictive performance with higher sensitivity than the qSOFA alone in predicting mortality in patients with complicated intra-abdominal infections and had a comparable predictive performance to that of the full SOFA score.

Prevalence of Malnutrition in Hospitalized Patients: a Multicenter Cross-sectional Study.

BACKGROUND: Malnutrition is associated with many adverse clinical outcomes. The present study aimed to identify the prevalence of malnutrition in hospitalized patients in Korea, evaluate the association between malnutrition and clinical outcomes, and ascertain the risk factors of malnutrition. METHODS: A multicenter cross-sectional study was performed with 300 patients recruited from among the patients admitted in 25 hospitals on January 6, 2014. Nutritional status was assessed by using the Subjective Global Assessment (SGA). Demographic characteristics and underlying diseases were compared according to nutritional status. Logistic regression analysis was performed to identify the risk factors of malnutrition. Clinical outcomes such as rate of admission in intensive care units, length of hospital stay, and survival rate were evaluated. RESULTS: The prevalence of malnutrition in the hospitalized patients was 22.0%. Old age (≥ 70 years), admission for medical treatment or diagnostic work-up, and underlying pulmonary or oncological disease were associated with malnutrition. Old age and admission for medical treatment or diagnostic work-up were identified to be risk factors of malnutrition in the multivariate analysis. Patients with malnutrition had longer hospital stay (SGA A = 7.63 ± 6.03 days, B = 9.02 ± 9.96 days, and C = 12.18 ± 7.24 days, P = 0.018) and lower 90-day survival rate (SGA A = 97.9%, B = 90.7%, and C = 58.3%, P < 0.001). CONCLUSION: Malnutrition was common in hospitalized patients, and resulted in longer hospitalization and associated lower survival rate. The rate of malnutrition tended to be higher when the patient was older than 70 years old or hospitalized for medical treatment or diagnostic work-up compared to elective surgery.

Hyporesponsiveness of natural killer cells and impaired inflammatory responses in critically ill patients.

BACKGROUND: To investigate natural killer (NK) cell activity, circulating cytokine level and peripheral blood mononuclear cell (PBMC) cytokine production status in critically ill patients. METHODS: Blood samples were collected <24 h after admission from 24 intensive care unit (ICU) patients and 24 age-, sex-, and body mass index (BMI)-matched healthy controls. Serum cytokine concentrations and cytokine production by PBMCs and lipopolysaccharide (LPS)-stimulated PBMCs were measured. RESULTS: The ICU group showed lower NK cell activity than the controls under all conditions and an absence of interferon (IFN)-γ. After adjusting for triglycerides, LDL- and HDL-cholesterol, and glucose, the ICU group exhibited lower serum levels of albumin and interleukin (IL)-12 and higher leukocyte counts and hs-CRP and IL-6 levels than the controls. Non-stimulated PBMCs from ICU patients secreted significantly greater amounts of IL-6 and IL-1ß than the controls; however, the production of IL-6, TNF-α and IL-1ß in response to LPS stimulation was significantly lower in the ICU group. CONCLUSIONS: Significant reductions in NK cell activity and serum IL-12 level, an absence of serum IFN-γ, and decreased cytokine production from LPS-stimulated PBMCs indicate the hyporesponsiveness of NK cells and an impaired early phase inflammatory response in critically ill patients (ClinicalTrials.gov NCT02565589 :). Retrospectively registered; October 1, 2015.

A nomogram predicting the need for abdominal and pelvic computed tomography in blunt trauma patients: A retrospective cohort study.

BACKGROUND: Abdominal and pelvic computed tomography (APCT) has become the preferred means for the initial evaluation of blunt trauma patients. However, computed tomography examination has some disadvantages, such as radiation exposure, the requirement for intravenous iodinated contrast medium, high cost, and time. We aimed to develop a nomogram to predict the need for APCT scanning after the primary survey of blunt trauma patients. MATERIALS AND METHODS: We conducted a retrospective observational cohort study at a single-center and reviewed medical records of 972 trauma patients admitted between January 2013 and June 2016. We enrolled 786 blunt trauma patients who had undergone APCT and were 16 years of age or older. A multivariate logistic regression model was used to determine independent predictors for trauma-related findings on APCT scans. A nomogram was constructed to predict injury on APCT scans based on each predictive factor. RESULTS: Of 786 patients, 355 (45%) patients had at least 1 injury on APCT scans. Results of multivariate logistic regression analysis showed that independent predictive factors of injuries on APCT scans were as follows: falls (≥3 m high); pain (abdominal, back, flank, or pelvic); positive peritoneal signs; abnormal findings on chest radiographs; abnormal findings on pelvic radiographs; and positive findings on focused assessment with ultrasonography for trauma. The nomogram was developed using these parameters. The area under a receiver operating characteristic curve of the multivariate model for discrimination was 0.865 (95% confidence interval, 0.840-0.892). The calibration plot showed good agreement between predicted and observed outcomes. The maximal Youden index was 0.59, corresponding to a cutoff value > 59 points, which was considered the optimal cutoff value for the probability that the injury would be detected on APCT scans. CONCLUSION: The nomogram, based on initial clinical findings in blunt trauma patients, will help clinicians be more selective in their use of APCT evaluations.