Dataset Information

External validation of a biomarker and clinical prediction model for hospital mortality in acute respiratory distress syndrome.

ABSTRACT: Mortality prediction in ARDS is important for prognostication and risk stratification. However, no prediction models have been independently validated. A combination of two biomarkers with age and APACHE III was superior in predicting mortality in the NHLBI ARDSNet ALVEOLI trial. We validated this prediction tool in two clinical trials and an observational cohort.The validation cohorts included 849 patients from the NHLBI ARDSNet Fluid and Catheter Treatment Trial (FACTT), 144 patients from a clinical trial of sivelestat for ARDS (STRIVE), and 545 ARDS patients from the VALID observational cohort study. To evaluate the performance of the prediction model, the area under the receiver operating characteristic curve (AUC), model discrimination, and calibration were assessed, and recalibration methods were applied.The biomarker/clinical prediction model performed well in all cohorts. Performance was better in the clinical trials with an AUC of 0.74 (95% CI 0.70-0.79) in FACTT, compared to 0.72 (95% CI 0.67-0.77) in VALID, a more heterogeneous observational cohort. The AUC was 0.73 (95% CI 0.70-0.76) when FACTT and VALID were combined.We validated a mortality prediction model for ARDS that includes age, APACHE III, surfactant protein D, and interleukin-8 in a variety of clinical settings. Although the model performance as measured by AUC was lower than in the original model derivation cohort, the biomarker/clinical model still performed well and may be useful for risk assessment for clinical trial enrollment, an issue of increasing importance as ARDS mortality declines, and better methods are needed for selection of the most severely ill patients for inclusion.

SUBMITTER: Zhao Z

PROVIDER: S-EPMC5978765 | biostudies-literature | 2017 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

External validation of a biomarker and clinical prediction model for hospital mortality in acute respiratory distress syndrome.

Zhao Zhiguo Z Wickersham Nancy N Kangelaris Kirsten N KN May Addison K AK Bernard Gordon R GR Matthay Michael A MA Calfee Carolyn S CS Koyama Tatsuki T Ware Lorraine B LB

Intensive care medicine 20170607 8

<h4>Purpose</h4>Mortality prediction in ARDS is important for prognostication and risk stratification. However, no prediction models have been independently validated. A combination of two biomarkers with age and APACHE III was superior in predicting mortality in the NHLBI ARDSNet ALVEOLI trial. We validated this prediction tool in two clinical trials and an observational cohort.<h4>Methods</h4>The validation cohorts included 849 patients from the NHLBI ARDSNet Fluid and Catheter Treatment Trial ...[more]

PMID: 28593401

Similar Datasets

Project description:ObjectivesAcute respiratory distress syndrome is frequently under recognized and associated with increased mortality. Previously, we developed a model that used machine learning and natural language processing of text from radiology reports to identify acute respiratory distress syndrome. The model showed improved performance in diagnosing acute respiratory distress syndrome when compared to a rule-based method. In this study, our objective was to externally validate the natural language processing model in patients from an independent hospital setting.DesignSecondary analysis of data across five prospective clinical studies.SettingAn urban, tertiary care, academic hospital.PatientsAdult patients admitted to the medical ICU and at-risk for acute respiratory distress syndrome.InterventionsNone.Measurements and main resultsThe natural language processing model was previously derived and internally validated in burn, trauma, and medical patients at Loyola University Medical Center. Two machine learning models were examined with the following text features from qualifying radiology reports: 1) word representations (n-grams) and 2) standardized clinical named entity mentions mapped from the National Library of Medicine Unified Medical Language System. The models were externally validated in a cohort of 235 patients at the University of Chicago Medicine, among which 110 (47%) were diagnosed with acute respiratory distress syndrome by expert annotation. During external validation, the n-gram model demonstrated good discrimination between acute respiratory distress syndrome and nonacute respiratory distress syndrome patients (C-statistic, 0.78; 95% CI, 0.72-0.84). The n-gram model had a higher discrimination for acute respiratory distress syndrome when compared with the standardized named entity model, although not statistically significant (C-statistic 0.78 vs 0.72; p = 0.09). The most important features in the model had good face validity for acute respiratory distress syndrome characteristics but differences in frequencies did occur between hospital settings.ConclusionsOur computable phenotype for acute respiratory distress syndrome had good discrimination in external validation and may be used by other health systems for case-identification. Discrepancies in feature representation are likely due to differences in characteristics of the patient cohorts.

Project description:BackgroundAcute respiratory distress syndrome (ARDS) is a common, but under-recognised, critical illness syndrome associated with high mortality. An important factor in its under-recognition is the variability in chest radiograph interpretation for ARDS. We sought to train a deep convolutional neural network (CNN) to detect ARDS findings on chest radiographs.MethodsCNNs were pretrained on 595 506 radiographs from two centres to identify common chest findings (eg, opacity and effusion), and then trained on 8072 radiographs annotated for ARDS by multiple physicians using various transfer learning approaches. The best performing CNN was tested on chest radiographs in an internal and external cohort, including a subset reviewed by six physicians, including a chest radiologist and physicians trained in intensive care medicine. Chest radiograph data were acquired from four US hospitals.FindingsIn an internal test set of 1560 chest radiographs from 455 patients with acute hypoxaemic respiratory failure, a CNN could detect ARDS with an area under the receiver operator characteristics curve (AUROC) of 0·92 (95% CI 0·89-0·94). In the subgroup of 413 images reviewed by at least six physicians, its AUROC was 0·93 (95% CI 0·88-0·96), sensitivity 83·0% (95% CI 74·0-91·1), and specificity 88·3% (95% CI 83·1-92·8). Among images with zero of six ARDS annotations (n=155), the median CNN probability was 11%, with six (4%) assigned a probability above 50%. Among images with six of six ARDS annotations (n=27), the median CNN probability was 91%, with two (7%) assigned a probability below 50%. In an external cohort of 958 chest radiographs from 431 patients with sepsis, the AUROC was 0·88 (95% CI 0·85-0·91). When radiographs annotated as equivocal were excluded, the AUROC was 0·93 (0·92-0·95).InterpretationA CNN can be trained to achieve expert physician-level performance in ARDS detection on chest radiographs. Further research is needed to evaluate the use of these algorithms to support real-time identification of ARDS patients to ensure fidelity with evidence-based care or to support ongoing ARDS research.FundingNational Institutes of Health, Department of Defense, and Department of Veterans Affairs.

Project description:OBJECTIVES:The original Pediatric Sepsis Biomarker Risk Model and revised (Pediatric Sepsis Biomarker Risk Model-II) biomarker-based risk prediction models have demonstrated utility for estimating baseline 28-day mortality risk in pediatric sepsis. Given the paucity of prediction tools in pediatric acute respiratory distress syndrome, and given the overlapping pathophysiology between sepsis and acute respiratory distress syndrome, we tested the utility of Pediatric Sepsis Biomarker Risk Model and Pediatric Sepsis Biomarker Risk Model-II for mortality prediction in a cohort of pediatric acute respiratory distress syndrome, with an a priori plan to revise the model if these existing models performed poorly. DESIGN:Prospective observational cohort study. SETTING:University affiliated PICU. PATIENTS:Mechanically ventilated children with acute respiratory distress syndrome. INTERVENTIONS:Blood collection within 24 hours of acute respiratory distress syndrome onset and biomarker measurements. MEASUREMENTS AND MAIN RESULTS:In 152 children with acute respiratory distress syndrome, Pediatric Sepsis Biomarker Risk Model performed poorly and Pediatric Sepsis Biomarker Risk Model-II performed modestly (areas under receiver operating characteristic curve of 0.61 and 0.76, respectively). Therefore, we randomly selected 80% of the cohort (n = 122) to rederive a risk prediction model for pediatric acute respiratory distress syndrome. We used classification and regression tree methodology, considering the Pediatric Sepsis Biomarker Risk Model biomarkers in addition to variables relevant to acute respiratory distress syndrome. The final model was comprised of three biomarkers and age, and more accurately estimated baseline mortality risk (area under receiver operating characteristic curve 0.85, p < 0.001 and p = 0.053 compared with Pediatric Sepsis Biomarker Risk Model and Pediatric Sepsis Biomarker Risk Model-II, respectively). The model was tested in the remaining 20% of subjects (n = 30) and demonstrated similar test characteristics. CONCLUSIONS:A validated, biomarker-based risk stratification tool designed for pediatric sepsis was adapted for use in pediatric acute respiratory distress syndrome. The newly derived Pediatric Acute Respiratory Distress Syndrome Biomarker Risk Model demonstrates good test characteristics internally and requires external validation in a larger cohort. Tools such as Pediatric Acute Respiratory Distress Syndrome Biomarker Risk Model have the potential to provide improved risk stratification and prognostic enrichment for future trials in pediatric acute respiratory distress syndrome.

Project description:ObjectivesPediatric acute respiratory distress syndrome is heterogeneous, with a paucity of risk stratification tools to assist with trial design. We aimed to develop and validate mortality prediction models for patients with pediatric acute respiratory distress syndrome.DesignLeveraging additional data collection from a preplanned ancillary study (Version 1) of the multinational Pediatric Acute Respiratory Distress syndrome Incidence and Epidemiology study, we identified predictors of mortality. Separate models were built for the entire Version 1 cohort, for the cohort excluding neurologic deaths, for intubated subjects, and for intubated subjects excluding neurologic deaths. Models were externally validated in a cohort of intubated pediatric acute respiratory distress syndrome patients from the Children's Hospital of Philadelphia.SettingThe derivation cohort represented 100 centers worldwide; the validation cohort was from Children's Hospital of Philadelphia.PatientsThere were 624 and 640 subjects in the derivation and validation cohorts, respectively.InterventionsNone.Measurements and main resultsThe model for the full cohort included immunocompromised status, Pediatric Logistic Organ Dysfunction 2 score, day 0 vasopressor-inotrope score and fluid balance, and PaO2/FIO2 6 hours after pediatric acute respiratory distress syndrome onset. This model had good discrimination (area under the receiver operating characteristic curve 0.82), calibration, and internal validation. Models excluding neurologic deaths, for intubated subjects, and for intubated subjects excluding neurologic deaths also demonstrated good discrimination (all area under the receiver operating characteristic curve ≥ 0.84) and calibration. In the validation cohort, models for intubated pediatric acute respiratory distress syndrome (including and excluding neurologic deaths) had excellent discrimination (both area under the receiver operating characteristic curve ≥ 0.85), but poor calibration. After revision, the model for all intubated subjects remained miscalibrated, whereas the model excluding neurologic deaths showed perfect calibration. Mortality models also stratified ventilator-free days at 28 days in both derivation and validation cohorts.ConclusionsWe describe predictive models for mortality in pediatric acute respiratory distress syndrome using readily available variables from day 0 of pediatric acute respiratory distress syndrome which outperform severity of illness scores and which demonstrate utility for composite outcomes such as ventilator-free days. Models can assist with risk stratification for clinical trials.

Project description:BackgroundMechanical power (MP) refers to the energy delivered by a ventilator to the respiratory system per unit of time. MP referenced to predicted body weight (PBW) or respiratory system compliance have better predictive value for mortality than MP alone in acute respiratory distress syndrome (ARDS). Our objective was to assess the potential impact of consecutive changes of MP on hospital mortality among ARDS patients receiving extracorporeal membrane oxygenation (ECMO).MethodsWe performed a retrospective analysis of patients with severe ARDS receiving ECMO in a tertiary care referral center in Taiwan between May 2006 and October 2015. Serial changes of MP during ECMO were recorded.ResultsA total of 152 patients with severe ARDS rescued with ECMO were analyzed. Overall hospital mortality was 53.3%. There were no significant differences between survivors and nonsurvivors in terms of baseline values of MP or other ventilator settings. Cox regression models demonstrated that mean MP alone, MP referenced to PBW, and MP referenced to compliance during the first 3 days of ECMO were all independently associated with hospital mortality. Higher MP referenced to compliance (HR 2.289 [95% CI 1.214-4.314], p = 0.010) was associated with a higher risk of death than MP itself (HR 1.060 [95% CI 1.018-1.104], p = 0.005) or MP referenced to PBW (HR 1.004 [95% CI 1.002-1.007], p < 0.001). The 90-day hospital mortality of patients with high MP (> 14.4 J/min) during the first 3 days of ECMO was significantly higher than that of patients with low MP (≦ 14.4 J/min) (70.7% vs. 46.8%, p = 0.004), and the 90-day hospital mortality of patients with high MP referenced to compliance (> 0.53 J/min/ml/cm H2O) during the first 3 days of ECMO was significantly higher than that of patients with low MP referenced to compliance (≦ 0.53 J/min/ml/cm H2O) (63.6% vs. 29.7%, p < 0.001).ConclusionsMP during the first 3 days of ECMO was the only ventilatory variable independently associated with 90-day hospital mortality, and MP referenced to compliance during ECMO was more predictive for mortality than was MP alone.

Project description:BackgroundThe lung microbiome maintains the homeostasis of the immune system within the lungs. In acute respiratory distress syndrome (ARDS), the lung microbiome is enriched with gut-derived bacteria; however, the specific microbiome associated with morbidity and mortality in patients with ARDS remains unclear. This study investigated the specific patterns of the lung microbiome that are correlated with mortality in ARDS patients.MethodsWe analyzed the lung microbiome from the bronchoalveolar lavage fluid (BALF) of patients with ARDS and control subjects. We measured the copy numbers of 16S rRNA and the serum and BALF cytokines (interleukin [IL]-6, IL-8, receptor for advanced glycation end products, and angiopoietin-2).ResultsWe analyzed 47 mechanically ventilated patients diagnosed with (n = 40) or without (n = 7; control) ARDS. The alpha diversity was significantly decreased in ARDS patients compared with that of the controls (6.24 vs. 8.07, P = 0.03). The 16S rRNA gene copy numbers tended to be increased in the ARDS group compared with the controls (3.83 × 106 vs. 1.01 × 105 copies/mL, P = 0.06). ARDS patients were subdivided into the hospital survivor (n = 24) and non-survivor groups (n = 16). Serum IL-6 levels were significantly higher in the non-survivors than in the survivors (567 vs. 214 pg/mL, P = 0.027). The 16S rRNA copy number was significantly correlated with serum IL-6 levels in non-survivors (r = 0.615, P < 0.05). The copy numbers and relative abundance of betaproteobacteria were significantly lower in the non-survivors than in the survivors (713 vs. 7812, P = 0.012; 1.22% vs. 0.08%, P = 0.02, respectively). Conversely, the copy numbers of Staphylococcus, Streptococcus and Enterobacteriaceae were significantly correlated with serum IL-6 levels in the non-survivors (r = 0.579, P < 0.05; r = 0.604, P < 0.05; r = 0.588, P < 0.05, respectively).ConclusionsThe lung bacterial burden tended to be increased, and the alpha diversity was significantly decreased in ARDS patients. The decreased Betaproteobacteria and increased Staphylococcus, Streptococcus and Enterobacteriaceae might represent a unique microbial community structure correlated with increased serum IL-6 and hospital mortality.Trial registrationThe institutional review boards of Hiroshima University (Trial registration: E-447-4, registered 16 October 2019) and Kyoto Prefectural University of Medicine (Trial registration: ERB-C-973, registered 19 October 2017) approved an opt-out method of informed consent.

Dataset Information

External validation of a biomarker and clinical prediction model for hospital mortality in acute respiratory distress syndrome.

Publications

External validation of a biomarker and clinical prediction model for hospital mortality in acute respiratory distress syndrome.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets