Decision support tool for early differential diagnosis of acute lung injury and cardiogenic pulmonary edema in medical critically ill patients

Learning using privileged information with logistic regression on acute respiratory distress syndrome detection

The advanced learning paradigm, learning using privileged information (LUPI), leverages information in training that is not present at the time of prediction. In this study, we developed privileged logistic regression (PLR) models under the LUPI paradigm to detect acute respiratory distress syndrome (ARDS), with mechanical ventilation variables or chest x-ray image features employed in the privileged domain and electronic health records in the base domain. In model training, the objective of privileged logistic regression was designed to incorporate data from the privileged domain and encourage knowledge transfer across the privileged and base domains. An asymptotic analysis was also performed, yielding sufficient conditions under which the addition of privileged information increases the rate of convergence in the proposed model. Results for ARDS detection show that PLR models achieve better classification performances than logistic regression models trained solely on the base domain, even when privileged information is partially available. Furthermore, PLR models demonstrate performance on par with or superior to state-of-the-art models under the LUPI paradigm. As the proposed models are effective, easy to interpret, and highly explainable, they are ideal for other clinical applications where privileged information is at least partially available.

Critical evaluation of established risk prediction models for acute respiratory distress syndrome in adult patients: A systematic review and meta-analysis

AIM

To assess the performance of validated prediction models for acute respiratory distress syndrome (ARDS) by systematic review and meta-analysis.

METHODS

Eight databases (Medline, CINAHL, Embase, The Cochrane Library, CNKI, WanFang Data, Sinomed, and VIP) were searched up to March 26, 2023. Studies developed and validated a prediction model for ARDS in adult patients were included. Items on study design, incidence, derivation methods, predictors, discrimination, and calibration were collected. The risk of bias was assessed by the Prediction model Risk of Bias Assessment Tool. Models with a reported area under the curve of the receiver operating characteristic (AUC) metric were analyzed.

RESULTS

A total of 25 studies were retrieved, including 48 unique prediction models. Discrimination was reported in all studies, with AUC ranging from 0.701 to 0.95. Emerged AUC value of the logistic regression model was 0.837 (95% CI: 0.814 to 0.859). Besides, the value in the ICU group was 0.856 (95% CI: 0.812 to 0.899), the acute pancreatitis group was 0.863 (95% CI: 0.844 to 0.882), and the postoperation group was 0.835 (95% CI: 0.808 to 0.861). In total, 24 of the included studies had a high risk of bias, which was mostly due to the improper methods in predictor screening (13/24), model calibration assessment (9/24), and dichotomization of continuous predictors (6/24).

CONCLUSIONS

This study shows that most prediction models for ARDS are at high risk of bias, and the discrimination ability of the model is excellent. Adherence to standardized guidelines for model development is necessary to derive a prediction model of value to clinicians.

Lung Ultrasound Signs to Diagnose and Discriminate Interstitial Syndromes in ICU Patients: A Diagnostic Accuracy Study in Two Cohorts

OBJECTIVES

To determine the diagnostic accuracy of lung ultrasound signs for both the diagnosis of interstitial syndrome and for the discrimination of noncardiogenic interstitial syndrome (NCIS) from cardiogenic pulmonary edema (CPE) in a mixed ICU population.

DESIGN

A prospective diagnostic accuracy study with derivation and validation cohorts.

SETTING

Three academic mixed ICUs in the Netherlands.

PATIENTS

Consecutive adult ICU patients that received a lung ultrasound examination.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULT

The reference standard was the diagnosis of interstitial syndrome (NCIS or CPE) or noninterstitial syndromes (other pulmonary diagnoses and no pulmonary diagnoses) based on full post-hoc clinical chart review except lung ultrasound. The index test was a lung ultrasound examination performed and scored by a researcher blinded to clinical information. A total of 101 patients were included in the derivation and 122 in validation cohort. In the derivation cohort, patients with interstitial syndrome ( n = 56) were reliably discriminated from other patients based on the presence of a B-pattern (defined as greater than or equal to 3 B-lines in one frame) with an accuracy of 94.7% (sensitivity, 90.9%; specificity, 91.1%). For discrimination of NCIS ( n = 29) from CPE ( n = 27), the presence of bilateral pleural line abnormalities (at least two: fragmented, thickened or irregular) had the highest diagnostic accuracy (94.6%; sensitivity, 89.3%; specificity, 100%). A diagnostic algorithm (Bedside Lung Ultrasound for Interstitial Syndrome Hierarchy protocol) using B-pattern and bilateral pleural abnormalities had an accuracy of 0.86 (95% CI, 0.77-0.95) for diagnosis and discrimination of interstitial syndromes. In the validation cohort, which included 122 patients with interstitial syndrome, bilateral pleural line abnormalities discriminated NCIS ( n = 98) from CPE ( n = 24) with a sensitivity of 31% (95% CI, 21-40%) and a specificity of 100% (95% CI, 86-100%).

CONCLUSIONS

Lung ultrasound can diagnose and discriminate interstitial syndromes in ICU patients with moderate-to-good accuracy. Pleural line abnormalities are highly specific for NCIS, but sensitivity is limited.

Assessment and diagnosis of chronic dyspnoea: a literature review

Dyspnoea or breathlessness is a common presenting symptom among patients attending primary care services. This review aimed to determine whether there are clinical tools that can be incorporated into a clinical decision support system for primary care for efficient and accurate diagnosis of causes of chronic dyspnoea. We searched MEDLINE, EMBASE and Google Scholar for all literature published between 1946 and 2020. Studies that evaluated a clinical algorithm for assessment of chronic dyspnoea in patients of any age group presenting to physicians with chronic dyspnoea were included. We identified 326 abstracts, 55 papers were reviewed, and eight included. A total 2026 patients aged between 20–80 years were included, 60% were women. The duration of dyspnoea was three weeks to 25 years. All studies undertook a stepwise or algorithmic approach to the assessment of dyspnoea. The results indicate that following history taking and physical examination, the first stage should include simply performed tests such as pulse oximetry, spirometry, and electrocardiography. If the patient remains undiagnosed, the second stage includes investigations such as chest x-ray, thyroid function tests, full blood count and NT-proBNP. In the third stage patients are referred for more advanced tests such as echocardiogram and thoracic CT. If dyspnoea remains unexplained, the fourth stage of assessment will require secondary care referral for more advanced diagnostic testing such as exercise tests. Utilising this proposed stepwise approach is expected to ascertain a cause for dyspnoea for 35% of the patients in stage 1, 83% by stage 3 and >90% of patients by stage 4.

Analytics with artificial intelligence to advance the treatment of acute respiratory distress syndrome

Artificial intelligence (AI) has found its way into clinical studies in the era of big data. Acute respiratory distress syndrome (ARDS) or acute lung injury (ALI) is a clinical syndrome that encompasses a heterogeneous population. Management of such heterogeneous patient population is a big challenge for clinicians. With accumulating ALI datasets being publicly available, more knowledge could be discovered with sophisticated analytics. We reviewed literatures with big data analytics to understand the role of AI for improving the caring of patients with ALI/ARDS. Many studies have utilized the electronic medical records (EMR) data for the identification and prognostication of ARDS patients. As increasing number of ARDS clinical trials data is open to public, secondary analysis on these combined datasets provide a powerful way of finding solution to clinical questions with a new perspective. AI techniques such as Classification and Regression Tree (CART) and artificial neural networks (ANN) have also been successfully used in the investigation of ARDS problems. Individualized treatment of ARDS could be implemented with a support from AI as we are now able to classify ARDS into many subphenotypes by unsupervised machine learning algorithms. Interestingly, these subphenotypes show different responses to a certain intervention. However, current analytics involving ARDS have not fully incorporated information from omics such as transcriptome, proteomics, daily activities and environmental conditions. AI technology is assisting us to interpret complex data of ARDS patients and enable us to further improve the management of ARDS patients in future with individual treatment plans.

Design and validation of a scoring model for differential diagnosis of diabetic nephropathy and nondiabetic renal diseases in type 2 diabetic patients

BACKGROUND

We aim to design a scoring model for differential diagnosis between diabetic nephropathy (DN) and nondiabetic renal disease (NDRD) in type 2 diabetic patients through a combination of clinical variables.

METHODS

A total of 170 patients with type 2 diabetes who underwent kidney biopsies were included and divided into three groups according to pathological findings: DN group (n = 46), MIX group (DN + NDRD, n = 54), NDRD group (n = 70). Clinical characteristics and laboratory data were collected and compared among groups. Variables with a significant statistical difference between DN and NDRD patients were analyzed by logistic regression to predict the presence of NDRD; then a scoring model was established based on the regression coefficient and further validated in an independent cohort of 67 patients prospectively.

RESULTS

On biopsy, 72.9% of patients had NDRD, and the most common pathological type was membranous nephropathy. The established scoring model for predicting NDRD included five predictors: age, systolic blood pressure, hemoglobin, duration of diabetes, and absence of diabetic retinopathy. The model demonstrated good discrimination and calibration (area under curve [AUC] 0.863, 95% CI, 0.800-0.925; Hosmer-Lemeshow [H-L] P = .062). Furthermore, high prediction accuracy (AUC = 0.900; 95% CI, 0.815-0.985) in the validation cohort proved the stability of the model.

CONCLUSIONS

We present a simple, robust scoring model for predicting the presence of NDRD with high accuracy (0.85) for the first time. This decision support tool provides a noninvasive method for differential diagnosis of DN and NDRD, which may help clinicians assess the risk-benefit ratio of kidney biopsy for type 2 diabetic patients with renal impairment.

A systematic review of diagnostic methods to differentiate acute lung injury/acute respiratory distress syndrome from cardiogenic pulmonary edema

BACKGROUND

Discriminating acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) from cardiogenic pulmonary edema (CPE) is often challenging. This systematic review examines studies using biomarkers or images to distinguish ALI/ARDS from CPE.

METHODS

Three investigators independently identified studies designed to distinguish ALI/ARDS from CPE in adults. Studies were identified from PubMed, and the Cochrane Central Register of Controlled Trials database until July 3, 2017.

RESULTS

Of 475 titles and abstracts screened, 38 full texts were selected for review, and we finally included 24 studies in this systematic review: 21 prospective observational studies, two retrospective observational studies, and one retrospective combined with prospective study. These studies compared various biomarkers to differentiate subjects with ALI/ARDS and in those with CPE, and 13 calculated the area under the receiver operator characteristic curve (AUC). The most commonly studied biomarker (four studies) was brain natriuretic peptide (BNP) and the discriminatory ability ranged from AUC 0.67-0.87 but the timing of measurement varied. Other potential biomarkers or tools have been reported, but only as single studies.

CONCLUSIONS

There were no identified biomarkers or tools with high-quality evidence for differentiating ALI/ARDS from CPE. Combining clinical criteria with validated biomarkers may improve the predictive accuracy.

External Validity of Electronic Sniffers for Automated Recognition of Acute Respiratory Distress Syndrome

INTRODUCTION

Automated electronic sniffers may be useful for early detection of acute respiratory distress syndrome (ARDS) for institution of treatment or clinical trial screening.

METHODS

In a prospective cohort of 2929 critically ill patients, we retrospectively applied published sniffer algorithms for automated detection of acute lung injury to assess their utility in diagnosis of ARDS in the first 4 ICU days. Radiographic full-text reports were searched for "edema" OR ("bilateral" AND "infiltrate") and a more detailed algorithm for descriptions consistent with ARDS. Patients were flagged as possible ARDS if a radiograph met search criteria and had a PaO₂/FiO₂ or SpO₂/FiO₂ of 300 or 315, respectively. Test characteristics of the electronic sniffers and clinical suspicion of ARDS were compared to a gold standard of 2-physician adjudicated ARDS.

RESULTS

Thirty percent of 2841 patients included in the analysis had gold standard diagnosis of ARDS. The simpler algorithm had sensitivity for ARDS of 78.9%, specificity of 52%, positive predictive value (PPV) of 41%, and negative predictive value (NPV) of 85.3% over the 4-day study period. The more detailed algorithm had sensitivity of 88.2%, specificity of 55.4%, PPV of 45.6%, and NPV of 91.7%. Both algorithms were more sensitive but less specific than clinician suspicion, which had sensitivity of 40.7%, specificity of 94.8%, PPV of 78.2%, and NPV of 77.7%.

CONCLUSIONS

Published electronic sniffer algorithms for ARDS may be useful automated screening tools for ARDS and improve on clinical recognition, but they are limited to screening rather than diagnosis because their specificity is poor.

Comparison of hospital mortality and long-term survival in patients with acute lung injury/ARDS vs cardiogenic pulmonary edema

BACKGROUND

Early differential diagnosis of acute lung injury (ALI) vs cardiogenic pulmonary edema (CPE) is important for selecting the most appropriate therapy, but the prognostic implications of this distinction have not been studied. Accurate prognostic information is essential for providing appropriate informed consent prior to initiation of mechanical ventilation.

METHODS

This is a long-term follow-up study of a previously established population-based cohort of critically ill adult patients with acute pulmonary edema admitted at a tertiary-care center during 2006 to 2009, in which post hoc expert review had established ALI vs CPE diagnosis. Using logistic and Cox regression, hospital mortality and long-term survival were compared in patients with ALI vs patients with CPE.

RESULTS

Of 328 patients (ALI = 155, CPE = 173), 240 patients (73%) died during a median follow-up of 160 days. After adjusting for confounders, patients with ALI were significantly more likely to die in the hospital (OR = 4.2, 95% CI = 2.3-7.8, n = 325, P < .001), but among hospital survivors the risk of death during follow-up was the same in both groups (hazard ratio = 1.13, 95% CI = 0.79-1.62, n = 229, P = .50). Independent predictors of mortality included age and APACHE (Acute Physiology and Chronic Health Evaluation) III score. Results were similar when restricting patients with ALI to the subset with ARDS (Berlin definition). In post hoc analyses, the mortality rate in hospital survivors compared with the general US population was significantly higher during the first 2 years but essentially converged by year five.

CONCLUSIONS

Although hospital mortality is higher in patients with ALI/ARDS compared with patients with CPE, long-term survival is similar in hospital survivors from both groups.

Decision support tool for differential diagnosis of Acute Respiratory Distress Syndrome (ARDS) vs Cardiogenic Pulmonary Edema (CPE): a prospective validation and meta-analysis

Introduction

We recently presented a prediction score providing decision support with the often-challenging early differential diagnosis of acute lung injury (ALI) vs cardiogenic pulmonary edema (CPE). To facilitate clinical adoption, our objective was to prospectively validate its performance in an independent cohort.

Methods

Over 9 months, adult patients consecutively admitted to any intensive care unit of a tertiary-care center developing acute pulmonary edema were identified in real-time using validated electronic surveillance. For eligible patients, predictors were abstracted from medical records within 48 hours of the alert. Post-hoc expert review blinded to the prediction score established gold standard diagnosis.

Results

Of 1,516 patients identified by electronic surveillance, data were abstracted for 249 patients (93% within 48 hours of disease onset), of which expert review (kappa 0.93) classified 72 as ALI, 73 as CPE and excluded 104 as “other”. With an area under the curve (AUC) of 0.81 (95% confidence interval =0.73 to 0.88) the prediction score showed similar discrimination as in prior cohorts (development AUC = 0.81, P = 0.91; retrospective validation AUC = 0.80, P = 0.92). Hosmer-Lemeshow test was significant (P = 0.01), but across eight previously defined score ranges probabilities of ALI vs CPE were the same as in the development cohort (P = 0.60). Results were the same when comparing acute respiratory distress syndrome (ARDS, Berlin definition) vs CPE.

Conclusion

The clinical prediction score reliably differentiates ARDS/ALI vs CPE. Pooled results provide precise estimates of the score’s performance which can be used to screen patient populations or to assess the probability of ALI/ARDS vs CPE in specific patients. The score may thus facilitate early inclusion into research studies and expedite prompt treatment.

Electronic supplementary material

The online version of this article (doi:10.1186/s13054-014-0659-x) contains supplementary material, which is available to authorized users.

Approach to the patient with the acute respiratory distress syndrome

Given the high incidence and mortality of acute respiratory distress syndrome (ARDS) in critically ill patients, every practitioner needs a bedside approach both for early identification of patients at risk for ARDS and for the appropriate evaluation of patients who meet the diagnostic criteria of ARDS. Recent advances such as the Lung Injury Prediction score, the Early Acute Lung Injury score, and validation of the SpO(2)/Fio(2) ratio for assessing the degree of hypoxemia are all practical tools to aid the practitioner in caring for patients at risk of ARDS.

Computerized decision support in adult and pediatric critical care

Computerized decision support (CDS) is the most advanced form of clinical decision support available and has evolved with innovative technologies to provide meaningful assistance to medical professionals. Critical care clinicians are in unique environments where vast amounts of data are collected on individual patients, and where expedient and accurate decisions are paramount to the delivery of quality healthcare. Many CDS tools are in use today among adult and pediatric intensive care units as diagnostic aides, safety alerts, computerized protocols, and automated recommendations for management. Some CDS use have significantly decreased adverse events and improved costs when carefully implemented and properly operated. CDS tools integrated into electronic health records are also valuable to researchers providing rapid identification of eligible patients, streamlining data-gathering and analysis, and providing cohorts for study of rare and chronic diseases through data-warehousing. Although the need for human judgment in the daily care of critically ill patients has limited the study and realization of meaningful improvements in overall patient outcomes, CDS tools continue to evolve and integrate into the daily workflow of clinicians, and will likely provide advancements over time. Through novel technologies, CDS tools have vast potential for progression and will significantly impact the field of critical care and clinical research in the future.

Mechanical ventilation and acute lung injury in emergency department patients with severe sepsis and septic shock: an observational study

OBJECTIVES

The objectives were to characterize the use of mechanical ventilation in the emergency department (ED), with respect to ventilator settings, monitoring, and titration and to determine the incidence of progression to acute lung injury (ALI) after admission, examining the influence of factors present in the ED on ALI progression.

METHODS

This was a retrospective, observational cohort study of mechanically ventilated patients with severe sepsis and septic shock (June 2005 to May 2010), presenting to an academic ED with an annual census of >95,000 patients. All patients in the study (n = 251) were analyzed for characterization of mechanical ventilation use in the ED. The primary outcome variable of interest was the incidence of ALI progression after intensive care unit (ICU) admission from the ED and risk factors present in the ED associated with this outcome. Secondary analyses included ALI present in the ED and clinical outcomes comparing all patients progressing to ALI versus no ALI. To assess predictors of progression to ALI, significant variables in univariable analyses at a p ≤ 0.10 level were candidates for inclusion in a bidirectional, stepwise, multivariable logistic regression analysis.

RESULTS

Lung-protective ventilation was used in 68 patients (27.1%) and did not differ based on ALI status. Delivered tidal volume was highly variable, with a median tidal volume delivered of 8.8 mL/kg ideal body weight (IBW; interquartile range [IQR] = 7.8 to 10.0) and a range of 5.2 to 14.6 mL/kg IBW. Sixty-nine patients (27.5%) in the entire cohort progressed to ALI after admission to the hospital, with a mean (±SD) onset of 2.1 (±1) days. Multivariable logistic regression analysis demonstrated that a higher body mass index (BMI), higher Sequential Organ Failure Assessment (SOFA) score, and ED vasopressor use were associated with progression to ALI. There was no association between ED ventilator settings and progression to ALI. Compared to patients who did not progress to ALI, patients progressing to ALI after admission from the ED had an increase in mechanical ventilator duration, vasopressor dependence, and hospital length of stay (LOS).

CONCLUSIONS

Lung-protective ventilation is uncommon in the ED, regardless of ALI status. Given the frequency of ALI in the ED, the progression shortly after ICU admission, and the clinical consequences of this syndrome, the effect of ED-based interventions aimed at reducing the sequelae of ALI should be investigated further.

Diagnostic performance of electronic syndromic surveillance systems in acute care: a systematic review

CONTEXT

Healthcare Electronic Syndromic Surveillance (ESS) is the systematic collection, analysis and interpretation of ongoing clinical data with subsequent dissemination of results, which aid clinical decision-making.

OBJECTIVE

To evaluate, classify and analyze the diagnostic performance, strengths and limitations of existing acute care ESS systems.

DATA SOURCES

All available to us studies in Ovid MEDLINE, Ovid EMBASE, CINAHL and Scopus databases, from as early as January 1972 through the first week of September 2012.

STUDY SELECTION

Prospective and retrospective trials, examining the diagnostic performance of inpatient ESS and providing objective diagnostic data including sensitivity, specificity, positive and negative predictive values.

DATA EXTRACTION

Two independent reviewers extracted diagnostic performance data on ESS systems, including clinical area, number of decision points, sensitivity and specificity. Positive and negative likelihood ratios were calculated for each healthcare ESS system. A likelihood matrix summarizing the various ESS systems performance was created.

RESULTS

The described search strategy yielded 1639 articles. Of these, 1497 were excluded on abstract information. After full text review, abstraction and arbitration with a third reviewer, 33 studies met inclusion criteria, reporting 102,611 ESS decision points. The yielded I2 was high (98.8%), precluding meta-analysis. Performance was variable, with sensitivities ranging from 21% -100% and specificities ranging from 5%-100%.

CONCLUSIONS

There is significant heterogeneity in the diagnostic performance of the available ESS implements in acute care, stemming from the wide spectrum of different clinical entities and ESS systems. Based on the results, we introduce a conceptual framework using a likelihood ratio matrix for evaluation and meaningful application of future, frontline clinical decision support systems.

β-Nicotinamide adenine dinucleotide attenuates lipopolysaccharide-induced inflammatory effects in a murine model of acute lung injury

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) occur in approximately 200,000 patients per year. Studies indicate that lung endothelium plays a significant role in ALI. The authors' recent in vitro studies demonstrate a novel mechanism of β-nicotinamide adenine dinucleotide (β-NAD)-induced protection against gram-positive (pneumolysin, PLY) and gram-negative (lipopolysaccharide, LPS) toxin-induced lung endothelial cell (EC) barrier dysfunction. The objective of the current study was to evaluate the protective effect of β-NAD against LPS-induced ALI in mice. C57BL/6J mice were randomly divided into 4 groups: vehicle, β-NAD, LPS, and LPS/β-NAD. After surgery, mice were allowed to recover for 24 hours. Evans blue dye-albumin (EBA) was given through the internal jugular vein 2 hours prior to the termination of the experiments. Upon sacrificing the animals, bronchoalveolar lavage fluid (BALF) was collected and the lungs were harvested. β-NAD treatment significantly attenuated the inflammatory response by means of reducing the accumulation of cells and protein in BALF, blunting the parenchymal neutrophil infiltration, and preventing capillary leak. In addition, the histological examination demonstrated decreased interstitial edema in the LPS/β-NAD specimens, as compared to the LPS-only specimens. The mRNA levels of the anti-inflammatory cytokines were up-regulated in the LPS group treated with β-NAD compared to the LPS-only-treated group. β-NAD treatment down-regulated the mRNA levels of the proinflammatory cytokines. These findings suggest that β-NAD could be investigated as a therapeutic option against bacterial toxin-induced lung inflammation and ALI in mice.