Federated machine learning for predicting acute kidney injury in critically ill patients: a multicenter study in Taiwan.

Purpose: To address the contentious data sharing across hospitals, this study adopted a novel approach, federated learning (FL), to establish an aggregate model for acute kidney injury (AKI) prediction in critically ill patients in Taiwan. Methods: This study used data from the Critical Care Database of Taichung Veterans General Hospital (TCVGH) from 2015 to 2020 and electrical medical records of the intensive care units (ICUs) between 2018 and 2020 of four referral centers in different areas across Taiwan. AKI prediction models were trained and validated thereupon. An FL-based prediction model across hospitals was then established. Results: The study included 16,732 ICU admissions from the TCVGH and 38,424 ICU admissions from the other four hospitals. The complete model with 60 features and the parsimonious model with 21 features demonstrated comparable accuracies using extreme gradient boosting, neural network (NN), and random forest, with an area under the receiver-operating characteristic (AUROC) curve of approximately 0.90. The Shapley Additive Explanations plot demonstrated that the selected features were the key clinical components of AKI for critically ill patients. The AUROC curve of the established parsimonious model for external validation at the four hospitals ranged from 0.760 to 0.865. NN-based FL slightly improved the model performance at the four centers. Conclusion: A reliable prediction model for AKI in ICU patients was developed with a lead time of 24 h, and it performed better when the novel FL platform across hospitals was implemented. Supplementary Information: The online version contains supplementary material available at 10.1007/s13755-023-00248-5.

Clinical and Microbiological Characteristics of Culture-Positive, Influenza-Associated Pulmonary Aspergillosis: A Single-Center Study in Southern Taiwan, 2016-2019.

This study delineated the characteristics of 24 (11.2%) culture-positive, influenza-associated pulmonary aspergillosis (IAPA) patients out of 215 patients with severe influenza during 2016-2019 in a medical center in southern Taiwan. Twenty (83.3%) patients did not have EORTC/MSG-defined host factors. The mean time from influenza diagnosis to Aspergillus growth was 4.4 days, and 20 (83.3%) developed IAPA within seven days after influenza diagnosis. All patients were treated in intensive care units and all but one (95.8%) received mechanical ventilation. Aspergillus tracheobronchitis was evident in 6 (31.6%) of 19 patients undergoing bronchoscopy. Positive galactomannan testing of either serum or bronchoalveolar lavage was noted in all patients. On computed tomography imaging, IAPA was characterized by peribronchial infiltrates, multiple nodules, and cavities superimposed on ground-glass opacities. Pure Aspergillus growth without bacterial co-isolation in culture was found in 17 (70.8%) patients. A. fumigatus (15, 62.5%), A. flavus (6, 25.0%), and A. terreus (4, 16.7%) were the major causative species. Three patients had mixed Aspergillus infections due to two species, and two had mixed azole-susceptible and azole-resistant A. fumigatus infection. All patients received voriconazole with an all-cause mortality of 41.6%. Of 14 survivors, the mean duration of antifungal use was 40.5 days. In conclusion, IAPA is an early and rapidly deteriorating complication following influenza that necessitates clinical vigilance and prompt diagnostic workup.

Invasive pulmonary aspergillosis is associated with cytomegalovirus viremia in critically ill patients - A retrospective cohort study.

BACKGROUND/PURPOSE: Cytomegalovirus (CMV) viremia is associated with a higher mortality rate and prolonged intensive care unit (ICU) stay for critically ill patients. CMV infection causes transient but substantial immunosuppression for transplant recipients, increasing risk of fungal infection. The association between CMV viremia and invasive pulmonary aspergillosis (IPA) for critically ill patients is still unknown. METHODS: We retrospectively analyzed patients received bronchoalveolar lavage (BAL), galactomannan test, influenza survey and blood CMV viral load test in ICUs of a university hospital between April 2017 and May 2020. Independent risks for IPA were analyzed by multivariable logistic regression. RESULTS: A total of 136 patients were included. Twenty-one patients had IPA, 48 patients had CMV viremia and 22 patients had influenza. In a multivariable logistic regression model, patients with CMV viremia or influenza had higher IPA risk (adjusted odds ratio, 3.98 and 8.72; 95% CI, 1.26-12.60 and 2.64-28.82; p value = 0.019 and <0.001, respectively.). Patients with detectable CMV in BAL fluid did not have higher IPA risk (crude odds ratio, 0.95; 95% CI, 0.33-2.79; p value = 0.933). After stratifying patients by CMV viral load, the IPA risk is higher for patients with higher viral loads. There is an additive synergistic effect on IPA risk between CMV viremia and influenza infection. CONCLUSION: For critically ill patients, CMV viremia is an independent risk factor of IPA. Patients with higher blood CMV viral loads have a higher risk of IPA. CMV viremia and influenza have an additive synergistic effect for IPA risk in critically ill patients.

Causes of death among dengue patients causes of death among hospitalized adults with dengue fever in Tainan, 2015: Emphasis on cardiac events and bacterial infections.

INTRODUCTION: The 2015 dengue outbreak in southern Taiwan caused substantial mortality. We analyzed the causes of death among these patients. MATERIALS AND METHODS: This retrospective study was conducted at a medical center in Tainan from August 2015 to December 2015. Dengue was diagnosed based on the detection of serum dengue NS1 antigen, IgM, or viral RNA in the blood. Causes of death were retrieved from chart reviews by three clinicians. RESULTS: There were 4488 cases of dengue in the study hospital, with an in-hospital fatality rate of 1.3% (60 cases). The mean age of the 60 fatal cases was 73 years, among whom 90% were aged ≥65 years. Twenty-eight (46.7%) patients died of severe dengue, and 29 (48.3%) deaths were possibly related to dengue. Of the latter, 24 (40%) died of secondary infections. Thirteen cardiac arrest events, including out-of-hospital (5 events) and in-hospital (8) cardiac arrests in the emergency department, occurred during the dengue epidemic. Seven (53.8%) patients did not receive medical aid before the event. Of the 40 deaths that occurred within one week after hospitalization, 60% died of severe dengue. In contrast, 50% of 20 deaths that occurred one week after hospitalization were related to hospital-acquired infections, mainly pneumonia. CONCLUSION: Of 60 fatal cases, with a predominance of elderly patients, deaths were related to severe dengue within the first week after admission and secondary infections thereafter. The absence of medical care before cardiac arrest events highlights the importance of health education for warning signs of dengue.

Severe Lung Infection and Septicemia Caused by Paludibacterium purpuratum-A Case Report and Evaluation of Bacterial Traits.

Paludibacterium species are gram stain-negative rods that are facultatively anaerobic; they have been isolated from wetland soil. Clinical infection caused by this genus is rarely reported. We report the case of an 84-year-old woman with chronic renal disease and hypertension who acquired P. purpuratum lung infection and septicemia in Southern Taiwan.

Respiratory viral infections in pragmatically selected adults in intensive care units.

Respiratory viruses can be detected in 18.3 to 48.9% of critically ill adults with severe respiratory tract infections (RTIs). The present study aims to assess the clinical significance of respiratory viruses in pragmatically selected adults in medical intensive care unit patients and to identify factors associated with viral respiratory viral tract infections (VRTIs). We conducted a prospective study on critically ill adults with suspected RTIs without recognized respiratory pathogens. Viral cultures with monoclonal antibody identification, in-house real-time polymerase chain reaction (PCR) for influenza virus, and FilmArray respiratory panel were used to detect viral pathogens. Multivariable logistic regression was applied to identify factors associated with VRTIs. Sixty-four (40.5%) of the included 158 critically ill adults had respiratory viruses detected in their respiratory specimens. The commonly detected viruses included influenza virus (20), followed by human rhinovirus/enterovirus (11), respiratory syncitial virus (9), human metapneumovirus (9), human parainfluenza viruses (8), human adenovirus (7), and human coronaviruses (2). The FilmArray respiratory panel detected respiratory viruses in 54 (34.6%) patients, but showed negative results for seven of 13 patients with influenza A/H3 infection. In the multivariable logistic regression model, patient characters associated with VRTIs included those aged < 65 years, household contact with individuals with upper RTI, the presence of fever, cough with sputum production, and sore throat. Respiratory viruses were not uncommonly detected in the pragmatically selected adults with critical illness. The application of multiplex PCR testing for respiratory viruses in selected patient population is a practical strategy, and the viral detection rate could be further improved by the patient characters recognized in this study.

A Review of Treatment of Coronavirus Disease 2019 (COVID-19): Therapeutic Repurposing and Unmet Clinical Needs.

For the initial phase of pandemic of coronavirus disease 2019 (COVID-19), repurposing drugs that in vitro inhibit severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) have been attempted with overlooked or overestimated efficacy owing to limited clinical evidence. Most early clinical trials have the defects of study design, small sample size, non-randomized design, or different timings of treatment initiation. However, well-designed studies on asymptomatic or mild, or pediatric cases of COVID-19 are scarce and desperately needed to meet the clinical need. However, a trend could be observed based on current clinical evidence. Remdesivir and favipiravir may shorten the recovery time; lopinavir/ritonavir does not demonstrate treatment efficacy in severe patients. Triple therapy of ribavirin, lopinavir, and interferon ß-1b showed early viral negative conversion, and the major effect may be related to interferon. Some small sample-size studies showed that interleukin-6 inhibitors may demonstrate clinical improvement; non-critical patients may benefit from convalescent plasma infusion in small sample-size studies; and the role of hydroxychloroquine or chloroquine in the treatment and prophylaxis of COVID-19 remains unclear. Combination therapy of traditional Chinese medicine with antiviral agents (ex. interferon, lopinavir, or arbidol) may alleviate inflammation in severe COVID-19 patients based on small sample-sized observational studies and experts' opinion. Most of the published studies included severe or critical patients with COVID-19. Combination therapy of antiviral agents and immune-modulating drugs is reasonable especially for those critical COVID-19 patients with cytokine release syndrome. Drugs to blunt cytokine release might not benefit for patients in the early stage with mild disease or the late stage with critical illness. Traditional Chinese medicine with antiviral effects on SARS-CoV-2 and immune-modulation is widely used for COVID-19 patients in China, and is worthy of further studies. In this review, we aim to highlight the available therapeutic options for COVID-19 based on current clinical evidence and encourage clinical trials specific for children and for patients with mild disease or at the early stage of COVID-19.

Implementation of a national bundle care program to reduce central line-associated bloodstream infections in intensive care units in Taiwan.

BACKGROUND/PURPOSE: This study assessed the effect of the central line bundle on the rate of central line-associated bloodstream infections (CLABSI) in intensive care units (ICUs) in Taiwan. METHODS: This national study was conducted in 27 ICUs with 404 beds total, including 15 medical ICUs, 11 surgical ICUs, and one mixed ICU. The study period was divided into two phases: a pre-intervention (between June 1, 2011 and October 31, 2011) and intervention phase (between December 1, 2011 and October 31, 2012). Outcome variables, including CLABSI rates (per 1000 catheter-days) and catheter utilization rates, were measured. RESULTS: The overall rate of CLABSI significantly decreased by 12.2% (p < 0.001) from 5.74 per 1000 catheter-days in the pre-intervention phase to 5.04 per 1000 catheter-days in the intervention phase. The catheter utilization rate decreased by 1.1% from 55.3% in the pre-intervention phase to 54.2% in the intervention phase. The decline in CLABSI varied significantly among hospital and ICU levels, except surgical ICUs (p = 0.59). CONCLUSIONS: Implementing a multidimensional central-line bundle significantly reduced the rates of CLABSI by 12.2% in nearly all participating ICUs, except surgical ICUs.

A Cohort Study of Adult Patients with Severe Dengue in Taiwanese Intensive Care Units: The Elderly and APTT Prolongation Matter for Prognosis.

BACKGROUND: There was a large dengue outbreak in Taiwan in 2015, in which the ages of the affected individuals were higher than those in other countries. The aim of this study was to explore the characteristics and prognostic factors for adults with severe dengue in intensive care units (ICUs). METHODS: All adults admitted to ICUs with dengue virus infection (DENV) at a medical center from July 1, 2015 to December 31, 2015 were enrolled. DENV was diagnosed by the presence of serum NS1 antigen, IgM antibodies to dengue virus, or dengue virus RNA by real-time reverse transcriptase polymerase chain reaction. Demographic data, clinical features, and lab data were collected, and a multivariate Cox model was used to identify the predictive factors for in-hospital mortality. RESULTS: Seventy-five patients admitted to ICUs with laboratory-confirmed DENV were enrolled (mean age 72.3±9.3 years). The most common comorbidities included hypertension (72.0%), diabetes (43.7%), and chronic kidney disease (22.7%). The in-hospital case fatality rate (CFR) was 41.3%. The patients who died were predominantly female, had higher disease severity at ICU admission, shorter ICU/hospital stay, longer initial activated partial thromboplastin time (APTT), and higher initial serum aspartate transaminase levels. Cardiac arrest before ICU admission (hazard ratio [HR]: 6.26 [1.91-20.54]), prolonged APTT (>48 seconds; HR: 3.91 [1.69-9.07]), and the presence of acute kidney injury on admission (HR: 2.48 [1.07-5.74]), were independently associated with in-hospital fatality in the Cox multivariate analysis. CONCLUSION: During the 2015 dengue outbreak in Taiwan, the patients with severe dengue in ICUs were characterized by old age, multiple comorbidities, and a high CFR. Organ failure (including cardiac failure, and renal failure) and coagulation disturbance (prolongation of initial APTT) were independent predictive factors for in-hospital fatality.

Community-onset bacteremia in kidney transplant recipients: The recipients fare well in terms of mortality and kidney injury.

BACKGROUND: Bloodstream infection is not uncommon in kidney transplant recipients (KTRs) and is associated with mortality, graft loss, and increased medical expenses. Whether these septic patients are more vulnerable to serious complications, resistant strains, or worse clinical outcomes than other patient groups in the community-onset settings remains undetermined. METHODS: A retrospective study was conducted at a medical center in southern Taiwan. Community-onset bacteremia in the KTRs and a control population at the emergency department were identified. Demographic data, clinical characteristics, bacteremic pathogens, antimicrobial resistance, and clinical outcomes were recorded. RESULTS: Forty-one bacteremic episodes in the KTRs and 82 episodes in control patients were studied. The KTR group had younger age, fewer malignancies, more urosepsis (61% vs. 22%, p = 0.004), and fewer biliary tract infections (0% vs. 13.4%, p = 0.018). Escherichia coli was the most commonly isolated pathogen in both the groups (51.2% and 41.5%, respectively). No Klebsiella pneumoniae bacteremia was noted in the KTRs, compared with 14 (17.1%) episodes in the control group (p = 0.010). Antimicrobial resistance profiles of bacteremic pathogens were similar (all p > 0.6). The KTRs with community-onset bacteremia did not have a worse outcome (in-hospital mortality rate: 2.4% vs. 10%, p = 0.172) nor more incomplete resolution of kidney injury after acute kidney injury events (21.1% vs. 25%, p > 0.99) than the control group. CONCLUSION: KTRs with community-onset bacteremia did not fare worse in terms of clinical outcome and kidney injury.